BEGIN:VCALENDAR VERSION:2.0 PRODID:-//pretalx//cfp.scipy.org//scipy2025 BEGIN:VTIMEZONE TZID:PST BEGIN:STANDARD DTSTART:20001029T030000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10;UNTIL=20061029T100000Z TZNAME:PST TZOFFSETFROM:-0700 TZOFFSETTO:-0800 END:STANDARD BEGIN:STANDARD DTSTART:20071104T030000 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:PST TZOFFSETFROM:-0700 TZOFFSETTO:-0800 END:STANDARD BEGIN:DAYLIGHT DTSTART:20000402T030000 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=4;UNTIL=20060402T110000Z TZNAME:PDT TZOFFSETFROM:-0800 TZOFFSETTO:-0700 END:DAYLIGHT BEGIN:DAYLIGHT DTSTART:20070311T030000 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:PDT TZOFFSETFROM:-0800 TZOFFSETTO:-0700 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT UID:pretalx-scipy2025-MP7C33@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:Despite its reputation for being slow\, Python is the leading l anguage of scientific computing\, which generally needs large-scale (fast) computations. This is because most scientific problems can be split into "metadata bookkeeping" and "number crunching\," where the latter is perfor med by array-oriented (vectorized) calls into precompiled routines.\n\nThi s tutorial is an introduction to array-oriented programming. We'll focus o n techniques that are equally useful in any array library\, with a particu lar focus on NumPy and JAX. You'll work in groups on four class projects: Conway's Game of Life using arrays\, iterative computations on arrays\, ju st-in-time (JIT) compilation for the Mandelbrot set\, and exploring data i n ragged arrays. DTSTAMP:20260417T070310Z LOCATION:Ballroom C SUMMARY:Thinking in arrays - Jim Pivarski\, Peter Fackeldey URL:https://cfp.scipy.org/scipy2025/talk/MP7C33/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-9Y38WQ@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:In this tutorial\, you will learn how to integrate Large Langua ge Models (LLMs) directly into Python programs as thoughtfully-designed co re components of the program rather than bolt-on additions. This hands-on session teaches design principles and practical techniques for incorporati ng LLM outputs into program control flow. We will use LlamaBot\, an open-s ource Python interface to LLMs\, focusing on local execution with local an d efficient models. DTSTAMP:20260417T070310Z LOCATION:Ballroom A SUMMARY:Building with LLMs Made Simple - Eric Ma URL:https://cfp.scipy.org/scipy2025/talk/9Y38WQ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-FXCRJW@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:This tutorial is an introduction to data visualization using th e popular Vega-Altair Python library. Vega-Altair provides a simple and ex pressive API\, enabling authors to rapidly create a wide range of interact ive charts. \n\nParticipants will explore the fundamentals of effective ch art design and gain hands-on experience building a variety of visualizatio ns using Vega-Altair's declarative API. Furthermore\, this tutorial will i ntroduce users to advanced topics such as data transformations and interac tion design. We will finish off by covering practical workflows such as in tegrating Vega-Altair into dashboarding systems\, publishing visualization s\, and creating reusable\, themed charting libraries. By the end of the s ession\, attendees will have the skills to leverage Vega-Altair for both r apid prototyping and production-ready visualizations in diverse environmen ts DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Vega-Altair: A Structured Way to Build Interactive Charts - Jon Mea se\, Dylan Wootton URL:https://cfp.scipy.org/scipy2025/talk/FXCRJW/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-KA7ZYR@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:As general purpose GPU programming has risen in popularity\, ma ny Python programmers have expressed a need to use this technology in thei r libraries and applications. They soon realize that the GPU landscape is vast and sometimes difficult to traverse for Python users. \n\nIn this t alk\, I will demystify the CUDA-enabled Accelerated Python landscape\, foc using on the advantages and disadvantages of popular libraries\, the commo n performance issues encountered\, and the best practices to getting the m ost out of your GPU. Topics include CuPy\, numba\, nvmath-python\, cuDF\, and cuML.\n\nThis talk is beginner-friendly\, but even the most seasoned programmer will gain insight into the Python GPU computing landscape. DTSTAMP:20260417T070310Z LOCATION:Room 316 SUMMARY:The Accelerated Python Developer's Toolbox - Katrina Riehl URL:https://cfp.scipy.org/scipy2025/talk/KA7ZYR/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-PLUNRN@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:The advancement of AI systems necessitates the need for interpr etability to address transparency\, biases\, risks\, and regulatory compli ance. The workshop teaches core techniques in interpretability\, including SHAP (game-theoretic feature attribution)\, GINI (decision tree impurity analysis)\, LIME (local surrogate models)\, and Permutation Importance (fe ature shuffling)\, which provide global and local explanations for model d ecisions. With hands-on building of interpretability tools and visualizati on techniques\, we explore how these methods enable bias detection and cli nical trust in healthcare diagnostics and develop the most effective strat egies in finance. These techniques are essential in building interpretable AI to address the challenges of the black-box models. DTSTAMP:20260417T070310Z LOCATION:Ballroom D SUMMARY:A Hands-on Tutorial towards building Explainable Machine Learning u sing SHAP\, GINI\, LIME\, and Permutation Importance - Dr. Debarshi Datta\ , Dr. Subhosit Ray URL:https://cfp.scipy.org/scipy2025/talk/PLUNRN/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-WSSAU7@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:This tutorial will explore GPU-accelerated clustering technique s using RAPIDS cuML\, optimizing algorithms like K-Means\, DBSCAN\, and HD BSCAN for large datasets. Traditional clustering methods struggle with sca lability\, but GPU acceleration significantly enhances performance and eff iciency. \n\nParticipants will learn to leverage dimensionality reduction techniques (PCA\, T-SNE\, UMAP) for better data visualization and apply hy perparameter tuning with Optuna and cuML. The session also includes real-w orld applications like topic modeling in NLP and customer segmentation. By the end\, attendees will be equipped to implement\, optimize\, and scale clustering algorithms effectively\, unlocking faster and more powerful ins ights in machine learning workflows. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Scaling Clustering for Big Data: Leveraging RAPIDS cuML - Allison D ing URL:https://cfp.scipy.org/scipy2025/talk/WSSAU7/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-DDV9NQ@cfp.scipy.org DTSTART;TZID=PST:20250707T080000 DTEND;TZID=PST:20250707T120000 DESCRIPTION:[Structured Query Language](https://duckdb.org/docs/sql/introdu ction.html) (or SQL for short) is a programming language to manage data in a database system and an essential part of any data engineer’s tool kit . In this tutorial\, you will learn how to use SQL to create databases\, t ables\, insert data into them and extract\, filter\, join data or make cal culations using queries. We will use DuckDB\, a new open source embedded i n-process database system that combines cutting edge database research wit h dataframe-inspired ease of use. DuckDB is only a pip install away (with zero dependencies)\, and runs right on your laptop. You will learn how to use DuckDB with your existing Python tools like Pandas\, Polars\, and Ibis to simplify and speed up your pipelines. Lastly\, you will learn how to u se SQL to create fast\, interactive data visualizations\, and how to teach your data how to fly and share it via the Cloud. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:All the SQL a Pythonista needs to know: an introduction to SQL and DataFrames with DuckDB - Guen Prawiroatmodjo\, Alex Monahan\, Jacob Matson URL:https://cfp.scipy.org/scipy2025/talk/DDV9NQ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LZWWA3@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:Spreadsheets are one of the most common ways to share and work with data which helpfully also works great in Python! In this tutorial\, w e will cover some of the basics and best pratice of consuming and producin g spreadsheets in Python as well as a deep dive into how to run Python dir ectly in your spreadsheets. We will introduce and dive deep into the new P ython in Excel features as well as the Anaconda Toolbox for Excel add-in. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Develop Pythonic spreadsheets running Python in and out of the grid - Sarah Kaiser\, Jim Kitchen URL:https://cfp.scipy.org/scipy2025/talk/LZWWA3/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-3YBVVH@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:**Ontologies** provide a powerful way to structure knowledge\, enable reasoning\, and support more meaningful queries compared to traditi onal data models. Recently\, interest in ontologies has resurged\, driven by advancements in language models\, reasoning capabilities\, and the grow ing adoption of platforms like Palantir Foundry. \n\nIn this hands- on tutorial\, participants will explore ontology development across multip le domains using a variety of Python-based tools such as `rdflib`\, `Owlre ady2`\, `PySpark`\, `Pandas`\, and `SciPy`. They will learn how ontologies facilitate semantic reasoning\, improve data interoperability\, and enhan ce query capabilities. \nAdditionally\, attendees will build a rudimenta ry reasoning engine to better understand inference mechanisms. \nThe tut orial emphasizes practical applications and comparisons with conventional data representations\, making it ideal for researchers\, data engineers\, and developers interested in knowledge representation and reasoning. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:The-Silmaril: Practice #ontology engineering with Python (and other languages). - Shaurya Agarwal URL:https://cfp.scipy.org/scipy2025/talk/3YBVVH/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-K3DQD9@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:Large Language Models (LLMs) have revolutionized natural langua ge processing\, but they come with limitations such as hallucinations and outdated knowledge. Retrieval-Augmented Generation (RAG) is a practical ap proach to mitigating these issues by integrating external knowledge retrie val into the LLM generation process. \n\nThis tutorial will introduce the core concepts of RAG\, walk through its key components\, and provide a han ds-on session for building a complete RAG pipeline. We will also cover adv anced techniques\, such as hybrid search\, re-ranking\, ensemble retrieval \, and benchmarking. By the end of this tutorial\, participants will be eq uipped with both the theoretical understanding and practical skills needed to build robust RAG pipeline. DTSTAMP:20260417T070310Z LOCATION:Ballroom A SUMMARY:Retrieval Augmented Generation (RAG) for LLMs - Sukhada Kulkarni\, Siyu Qian\, Antoni Liria Sala\, Xinling Luo URL:https://cfp.scipy.org/scipy2025/talk/K3DQD9/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-MHNTAD@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:[PyVista](https://github.com/pyvista/pyvista) is a general purp ose 3D visualization library used for over 2000+ open source projects for the visualization of everything from [computer aided engineering and geoph ysics to volcanoes and digital artwork](https://dev.pyvista.org/getting-st arted/external_examples.html).\n\nPyVista exposes a Pythonic API to the [V isualization Toolkit (VTK)](http://www.vtk.org) to provide tooling that is immediately usable without any prior knowledge of VTK and is being built as the 3D equivalent of Matplotlib\, with plugins to Jupyter to enable vis ualization of 3D data using both server- and client-side rendering. DTSTAMP:20260417T070310Z LOCATION:Ballroom D SUMMARY:3D Visualization with PyVista - Alexander Kaszynski\, Tetsuo Koyama \, Bane Sullivan URL:https://cfp.scipy.org/scipy2025/talk/MHNTAD/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-ZAKQHP@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:Pandas and scikit-learn have become staples in the machine lear ning toolkit for processing and modeling tabular data in Python. However\, when data size scales up\, these tools become slow or run out of memory. Ibis provides a unified\, Pythonic\, dataframe-like interface to 20+ execu tion backends\, including dataframe libraries\, databases\, and analytics engines. Ibis enables users to leverage these powerful tools without rewri ting their data engineering code (or learning SQL). IbisML extends the ben efits of using Ibis to the ML workflow by letting users preprocess their d ata at scale on any Ibis-supported backend.\n\nIn this tutorial\, you'll b uild an end-to-end machine learning project to predict the live win probab ility after each move during chess games. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Building machine learning pipelines that scale: a case study using Ibis and IbisML - Deepyaman Datta\, Anjali Datta URL:https://cfp.scipy.org/scipy2025/talk/ZAKQHP/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-GDN8PN@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:Scientific researchers need reproducible software environments for complex applications that can run across heterogeneous computing platf orms. Modern open source tools\, like [`pixi`](https://pixi.sh/)\, provide automatic reproducibility solutions for all dependencies while providing a high level interface well suited for researchers.\n\nThis tutorial will provide a practical introduction to using `pixi` to easily create scientif ic and AI/ML environments that benefit from hardware acceleration\, across multiple machines and platforms. The focus will be on applications using the PyTorch and JAX Python machine learning libraries with CUDA enabled\, as well as deploying these environments to production settings in Linux co ntainer images. DTSTAMP:20260417T070310Z LOCATION:Ballroom C SUMMARY:Reproducible Machine Learning Workflows for Scientists with pixi - John Kirkham\, Matthew Feickert\, Ruben Arts URL:https://cfp.scipy.org/scipy2025/talk/GDN8PN/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-TPGZFY@cfp.scipy.org DTSTART;TZID=PST:20250707T133000 DTEND;TZID=PST:20250707T173000 DESCRIPTION:Working with data can be challenging: it often doesn’t come i n the best format for analysis\, and understanding it well enough to extra ct insights requires both time and the skills to filter\, aggregate\, resh ape\, and visualize it. This session will equip you with the knowledge you need to effectively use pandas – a powerful library for data analysis i n Python – to make this process easier.\n\nPandas makes it possible to w ork with tabular data and perform all parts of the analysis from collectio n and manipulation through aggregation and visualization. While most of th is session focuses on pandas\, during our discussion of visualization\, we will also introduce at a high level Matplotlib (the library that pandas u ses for its visualization features\, which when used directly makes it pos sible to create custom layouts\, add annotations\, etc.) and Seaborn (anot her plotting library\, which features additional plot types and the abilit y to visualize long-format data). DTSTAMP:20260417T070310Z LOCATION:Room 316 SUMMARY:Introduction to Data Analysis Using Pandas - Stefanie Molin URL:https://cfp.scipy.org/scipy2025/talk/TPGZFY/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-WHKNQJ@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:The rapid expansion of the geospatial industry and accompanying increase in availability of geospatial data\, presents unique opportuniti es and challenges in data science. As the need for skilled data scientists increases\, the ability to manipulate and interpret this data becomes cru cial. This workshop introduces the essentials of geospatial data manipulat ion and data visualisation\, emphasizing hands-on techniques to transform\ , analyze and visualise diverse datasets effectively.\n\n\nThroughout the workshop\, attendees will explore the extensive ecosystem of geospatial Py thon libraries. Key tools include GeoPandas\, Shapely and Cartopy for vect or data\, GDAL\, Rasterio and rioxarray for raster data and participants w ill also learn to integrate these with popular plotting libraries such as Matplotlib\, Bokeh\, and Plotly for visualizations.\n\n\nThis tutorial wil l cover three primary topics: visualizing geospatial shapes\, managing ras ter datasets\, and synthesizing multiple data types into unified visual re presentations. Each section will incorporate data manipulation exercises t o ensure attendees not only visualize but also deeply understand geospatia l data.\n\n\nTargeting both beginners and advanced practitioners\, the wor kshop will employ real-world examples to guide participants through the ne cessary steps to produce striking and informative geospatial visualization s. By the end\, attendees will be equipped with the knowledge to leverage advanced data science techniques in their geospatial projects\, making the m proficient in both the analysis and communication of spatial information . DTSTAMP:20260417T070310Z LOCATION:Ballroom D SUMMARY:Geospatial data visualisation in Python - Adam Symington URL:https://cfp.scipy.org/scipy2025/talk/WHKNQJ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RYTBM8@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:**TL\;DR**\nLearn how to turn your Python functions into intera ctive web applications using open-source tools. By the end\, each of us wi ll have deployed a portfolio (or store) with multiple web applications and learned how to reproduce it easily later on.\n\n**Tell me more**\nWork no t shown is work lost. Many excellent scientists and engineers are not alwa ys adept at showcasing their work. This results in many interesting scient ific ideas that have never been brought to light.\n\nHowever\, using today 's tools\, one no longer has to leave the Python ecosystem to create class y\, complete prototypes using modern data visualization and web developmen t tools. With over five years of experience building and presenting data s olutions at huge science companies\, we show it doesn't have to be challen ging. We provide a walkthrough of the primary web application frameworks a nd showcase Fast Dash\, an open-source Python library that we built to add ress specific prototyping needs.\n\nThis tutorial is designed for all data professionals who value the ability to quickly convert their scientific c ode into web applications. Participants will learn about the leading frame works\, their strengths and limitations\, and a decision flowchart for pic king the best one for a given task. We will go through some day-to-day app lications and hands-on Python coding throughout the session. Whether you b ring your use-cases and datasets\, or pick from our suggestions\, you'll h ave a reproducible portfolio (app store) of deployed web applications by t he end! DTSTAMP:20260417T070310Z LOCATION:Ballroom C SUMMARY:Show your work: Tutorial on building and hosting web applications - Kedar Dabhadkar\, Archit Datar URL:https://cfp.scipy.org/scipy2025/talk/RYTBM8/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-FNUDXC@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:Cloud-optimized (CO) data formats are designed to efficiently s tore and access data directly from cloud storage without needing to downlo ad the entire dataset.\nThese formats enable faster data retrieval\, scala bility\, and cost-effectiveness by allowing users to fetch only the necess ary subsets of data.\nThey also allow for efficient parallel data processi ng using on-the-fly partitioning\, which can considerably accelerate data management operations.\nIn this sense\, cloud-optimized data is a nice fit for data-parallel jobs using serverless.\nFaaS provides a data-driven sca lable and cost-efficient experience\, with practically no management burde n.\nEach serverless function will read and process a small portion of the cloud-optimized dataset\, being read in parallel directly from object stor age\, significantly increasing the speedup.\n\nIn this talk\, you will lea rn how to process cloud-optimized data formats in Python using the Lithops toolkit.\n[Lithops](https://github.com/lithops-cloud/lithops) is a server less data processing toolkit that is specially designed to process data fr om Cloud Object Storage using Serverless functions.\nWe will also demonstr ate the [Dataplug library](https://github.com/CLOUDLAB-URV/dataplug) that enables Cloud Optimized data managament of scientific settings such as gen omics\, metabolomics\, or geospatial data. We will show different data pro cessing pipelines\nin the Cloud that demonstrate the benefits of cloud-opt imized data management. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Processing Cloud-optimized data in Python with Serverless Functions (Lithops\, Dataplug) - Universitat Rovira i Virgili (Pedro Garcia Lopez)\ , Enrique Molina Giménez URL:https://cfp.scipy.org/scipy2025/talk/FNUDXC/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-MDJVGA@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:With cameras in everything from microscopes to telescopes to sa tellites\, scientists produce image data in countless formats\, shapes\, s izes\, and dimensions. Python provides a rich ecosystem of libraries to ma ke sense of them. napari is a Python library for multidimensional image vi sualization\, but it does double duty as a standalone application that can be easily extended with GUI tools for analysis\, visualization\, and anno tation. In this tutorial\, we'll start with the basics of image visualizat ion and analysis in Python\, then show how to extend the napari user inter face to make analysis workflows as easy as pushing a button\, and finally show how to share these extensions as *plugins*\, which can be easily inst alled by users and collaborators. If you work with images (particularly mu ltidimensional images)\, and especially if you work with scientists who ma y not be comfortable with Python\, this tutorial might be for you! DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Create custom image visualization and analysis tools with napari - Tim Monko\, Draga Doncila Pop\, Peter Sobolewski URL:https://cfp.scipy.org/scipy2025/talk/MDJVGA/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-7DDV7V@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:Through the use of NetworkX's API\, tutorial participants will learn about the basics of graph theory and its use in applied network scie nce. Starting with a computationally-oriented definition of a graph and it s associated methods\, we will progress through the following concepts: pa th and structure finding\, visualization\, and graph storage on disk. We w ill also offer tutorial participants the option of one advanced topic over view\, including the use of graphs alongside LLMs for knowledge retrieval\ , scalable alternatives to NetworkX including cuGraph\, and the use of lin ear algebraic translation of graph problems to speed up computations. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Network Analysis Made Simple - Eric Ma URL:https://cfp.scipy.org/scipy2025/talk/7DDV7V/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-Z3VBWR@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:Python packaging can be overwhelming. However\, a trusted\, com munity-vetted workflow can make it easier. In this hands-on workshop\, you ’ll learn a tested approach developed by the pyOpenSci community and vet ted by Python packaging maintainers. You’ll create an installable\, main tainable\, and citable package using a quickstart template. You’ll also receive step-by-step guidance on publishing to TestPyPI (and resources for conda-forge\, and adding a DOI with Zenodo). If you can’t install softw are on your laptop\, you can use GitHub Codespaces to participate in the w orkshop. Join us to package your Python code confidently and to access ong oing support in our community beyond the workshop. DTSTAMP:20260417T070310Z LOCATION:Room 316 SUMMARY:Create Your First Python Package: Make Your Python Code Easier to S hare and Use - Carol Willing\, Jeremiah Paige\, Tetsuo Koyama\, Leah Wasse r\, Inessa Pawson URL:https://cfp.scipy.org/scipy2025/talk/Z3VBWR/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-XUYKZZ@cfp.scipy.org DTSTART;TZID=PST:20250708T080000 DTEND;TZID=PST:20250708T120000 DESCRIPTION:This hands-on tutorial will guide participants through building an end-to-end AI agent that translates natural language questions into SQ L queries\, validates and executes them on live databases\, and returns ac curate responses. Participants will build a system that intelligently rout es between a specialized SQL agent and a ReAct chat agent\, implementing R AG for query similarity matching\, comprehensive safety validation\, and h uman-in-the-loop confirmation. By the end of this 4-hour session\, attende es will have created a powerful and extensible system they can adapt to th eir own data sources. DTSTAMP:20260417T070310Z LOCATION:Ballroom A SUMMARY:Building an AI Agent for Natural Language to SQL Query Execution on Live Databases - Cainã Max Couto da Silva URL:https://cfp.scipy.org/scipy2025/talk/XUYKZZ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-HEHW8W@cfp.scipy.org DTSTART;TZID=PST:20250708T133000 DTEND;TZID=PST:20250708T173000 DESCRIPTION:Shiny is a framework for building web applications and data das hboards in Python.\nIn this workshop\,\nyou will see how the basic buildin g blocks of shiny can be extended to create\nyour own scalable production- ready python applications.\n\nIn particular\, this workshop covers:\n\n- O verview of the basic building blocks of a Shiny for Python application\n- How to refactor applications into shiny modules\n- How to write tests for your shiny application\n- Deploy and share your application\n\nAt the end of this course you will be able to:\n\n- Build a Shiny app in Python\n- Re factor your reactive logic into Shiny Modules\n- Identify when to write Sh iny modules\n- Write unit tests and end-to-end tests for your shiny applic ation\n- Deploy and share your application (for free!) DTSTAMP:20260417T070310Z LOCATION:Ballroom D SUMMARY:Shiny for Python: Building Production-Ready Dashboards in Python - Daniel Chen URL:https://cfp.scipy.org/scipy2025/talk/HEHW8W/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-8HD89Q@cfp.scipy.org DTSTART;TZID=PST:20250708T133000 DTEND;TZID=PST:20250708T173000 DESCRIPTION:Artificial intelligence has been successfully applied to bioima ge understanding and achieved significative results in the last decade. Ad vances in imaging technologies have also allowed the acquisition of higher resolution images. That has increased not only the magnification at what images are captured\, but the size of the acquired images as well. This co mprises a challenge for deep learning inference in large-scale images\, si nce these methods are commonly used in relatively small regions rather tha n whole images. This workshop presents techniques to scale-up inference of deep learning models to large-scale image data with help of *Dask* for pa rallelization in Python. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Scaling-up deep learning inference to large-scale bioimage data - F ernando Cervantes Sanchez\, Peter Sobolewski URL:https://cfp.scipy.org/scipy2025/talk/8HD89Q/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-87VTR7@cfp.scipy.org DTSTART;TZID=PST:20250708T133000 DTEND;TZID=PST:20250708T173000 DESCRIPTION:Maintaining code quality can be challenging\, no matter the siz e of your project or number of contributors. Different team members may ha ve different opinions on code styling and preferences for code structure\, while solo contributors might find themselves spending a considerable amo unt of time making sure the code conforms to accepted conventions. However \, manually inspecting and fixing issues in files is both tedious and erro r-prone. As such\, computers are much more suited to this task than humans . Pre-commit hooks are a great way to have a computer handle this for you. \n\nPre-commit hooks are code checks that run whenever you attempt to comm it your changes with Git. They can detect and\, in some cases\, automatica lly correct code-quality issues *before* they make it to your codebase. In this tutorial\, you will learn how to install and configure pre-commit ho oks for your repository to ensure that only code that passes your checks m akes it into your code base. We will also explore how to build custom pre- commit hooks for novel use cases. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:(Pre-)Commit to Better Code - Stefanie Molin URL:https://cfp.scipy.org/scipy2025/talk/87VTR7/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RPN9U9@cfp.scipy.org DTSTART;TZID=PST:20250708T133000 DTEND;TZID=PST:20250708T173000 DESCRIPTION:As data science continues to evolve\, the ever-growing size of datasets poses significant computational challenges. Traditional CPU-based processing often struggles to keep pace with the demands of data science workflows. Accelerated computing with GPUs offers a solution by enabling m assive parallelism and significantly reducing processing times for data-he avy tasks. In this session\, we will explore GPU computing architecture\, how it differs from CPUs\, and why it is particularly well-suited for data science workloads. This hands-on lab will dive into the different approac hes to GPU programming\, from low-level CUDA coding to high-level Python l ibraries within RAPIDS such as\, CuPy\, cuDF\, cuGraph\, and cuML. DTSTAMP:20260417T070310Z LOCATION:Room 316 SUMMARY:Bring Accelerated Computing to Data Science in Python - Kevin Lee URL:https://cfp.scipy.org/scipy2025/talk/RPN9U9/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-A8D9Z7@cfp.scipy.org DTSTART;TZID=PST:20250708T133000 DTEND;TZID=PST:20250708T173000 DESCRIPTION:Xarray provides data structures for multi-dimensional labeled a rrays and a toolkit for scalable data analysis on large\, complex datasets . Many real-world datasets often have hierarchical or heterogeneous struct ure\, and are best organized through groups of related data arrays. Throug h xarray.DataTree\, the xarray data model now supports opening datasets wi th a hierarchical structure of groups\, such as HDF5 files and Zarr stores . This expanded data model is now general enough to manage data across dif ferent scientific disciplines\, including geosciences and biosciences. Thi s hands-on tutorial focuses on intermediate and advanced workflows using x array to analyze real-world hierarchical data. DTSTAMP:20260417T070310Z LOCATION:Ballroom C SUMMARY:Hierarchical Data Analysis with Xarray DataTree & Zarr - Deepak Che rian\, Ian Hunt-Isaak\, Eniola Awowale\, Tom Nicholas\, Scott Henderson\, Justus Magin\, Joe Hamman\, Negin Sobhani URL:https://cfp.scipy.org/scipy2025/talk/A8D9Z7/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-7UJEBD@cfp.scipy.org DTSTART;TZID=PST:20250708T133000 DTEND;TZID=PST:20250708T173000 DESCRIPTION:This workshop is designed to equip software engineers with the skills to build and iterate on generative AI-powered applications. Partici pants will explore key components of the AI software development lifecycle through first principles thinking\, including prompt engineering\, monito ring\, evaluations\, and handling non-determinism. The session focuses on using multimodal AI models to build applications\, such as querying PDFs\, while providing insights into the engineering challenges unique to AI sys tems. By the end of the workshop\, participants will know how to build a P DF-querying app\, but all techniques learned will be generalizable for bui lding a variety of generative AI applications.\n\nIf you're a data scienti st\, machine learning practitioner\, or AI enthusiast\, this workshop can also be valuable for learning about the software engineering aspects of AI applications\, such as lifecycle management\, iterative development\, and monitoring\, which are critical for production-level AI systems. DTSTAMP:20260417T070310Z LOCATION:Ballroom A SUMMARY:Building LLM-Powered Applications for Data Scientists and Software Engineers - Stefan Krawczyk\, hugo bowne-anderson URL:https://cfp.scipy.org/scipy2025/talk/7UJEBD/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-BP7ADY@cfp.scipy.org DTSTART;TZID=PST:20250709T091500 DTEND;TZID=PST:20250709T100000 DESCRIPTION:Scientific Python is not only at the heart of discovery and adv ancement\, but also infrastructure. This talk will provide a perspective o n how open-source Python tools that are already powering real-world impact across the sciences are also supportive of public institutions and critic al public data infrastructure. Drawing on her previous experience leading policy efforts in the Department of Energy as well as her experience in op en-source scientific computing\, Katy will highlight the indispensable rol e of transparency\, reproducibility\, and community in high-stakes domains . This talk invites the SciPy community to recognize its unique strengths and to amplify their impact by contributing to the public good through tec hnically excellent\, civic-minded development. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:What We Maintain\, We Defend - Hon. Kathryn D. Huff\, PhD URL:https://cfp.scipy.org/scipy2025/talk/BP7ADY/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-DM3QPX@cfp.scipy.org DTSTART;TZID=PST:20250709T104500 DTEND;TZID=PST:20250709T111500 DESCRIPTION:A lot of data scientists use UMAP to help them quickly visualiz e and explore complex datasets. This could be exploring large unstructured datasets via neural embeddings\, or working on LLM explainability by mapp ing out Sparse Autoencoder features. Making the visualizations good enough \, and compelling enough\, to present to end users is much harder. However \, if done right a good UMAP plot can be a powerful communication tool\, o r a rich interactive experience that draws users in. Attendees will come a way with a sense of what is possible\, and an introduction to open source tools that can make it easy. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:DataMapPlot: Rich Tools for UMAP Visualizations - Leland McInnes URL:https://cfp.scipy.org/scipy2025/talk/DM3QPX/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-PRSN9R@cfp.scipy.org DTSTART;TZID=PST:20250709T104500 DTEND;TZID=PST:20250709T111500 DESCRIPTION:The practice of data science in genomics and computational biol ogy is fraught with friction. This is largely due to a tight coupling of b ioinformatic tools to file input/output. While omic data is specialized an d the storage formats for high-throughput sequencing and related data are often standardized\, the adoption of emerging open standards not tied to b ioinformatics can help better integrate bioinformatic workflows into the w ider data science\, visualization\, and AI/ML ecosystems. Here\, we presen t two bridge libraries as short vignettes for composable bioinformatics. F irst\, we present Anywidget\, an architecture and toolkit based on modern web standards for sharing interactive widgets across all Jupyter-compatibl e runtimes\, including JupyterLab\, Google Colab\, VSCode\, and more. Seco nd\, we present Oxbow\, a Rust and Python-based adapter library that unifi es access to common genomic data formats by efficiently transforming queri es into Apache Arrow\, a standard in-memory columnar representation for ta bular data analytics. Together\, we demonstrate the composition of these l ibraries to build a custom connected genomic analysis and visualization en vironments. We propose that components such as these\, which leverage scie ntific domain-agnostic standards to unbundle specialized file manipulation \, analytics\, and web interactivity\, can serve as reusable building bloc ks for composing flexible genomic data analysis and machine learning workf lows as well as systems for exploratory data analysis and visualization. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Accelerating Genomic Data Science and AI/ML with Composability - Tr evor Manz\, Nezar Abdennur URL:https://cfp.scipy.org/scipy2025/talk/PRSN9R/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-NCLV7B@cfp.scipy.org DTSTART;TZID=PST:20250709T104500 DTEND;TZID=PST:20250709T111500 DESCRIPTION:Matplotlib is already a favorite plotting library for creating static data visualizations in Python.\nHere\, we discuss the development o f a new `DataContainer` interface and accompanying transformation pipeline which enable easier dynamic data visualization in Matplotlib.\nThis impro ves the experience of plotting pure functions\, automatically recomputing when you pan and zoom.\nData containers can ingest data from a variety of sources\, including structured data such as Pandas Dataframes or Xarrays\, up to live updating data from web services or databases.\nThe flexible tr ansformation pipeline allows for control over how your data is encoded int o a plot. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Dynamic Data with Matplotlib - Kyle Sunden URL:https://cfp.scipy.org/scipy2025/talk/NCLV7B/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-BJL3U3@cfp.scipy.org DTSTART;TZID=PST:20250709T104500 DTEND;TZID=PST:20250709T111500 DESCRIPTION:Climate models generate *a lot* of data - and this can make it hard for researchers to efficiently access and use the data they need. The solutions of yesteryear include standardised file structures\, sqlite dat abases\, and just knowing where to look. All of these work - to varying de grees - but can leave new users scratching their heads. In this talk\, I'l l outline how ACCESS-NRI built tooling around Intake and Intake-ESM to mak e it easy for climate researchers to access available data\, share their o wn\, and avoid writing the custom scripts over and over to work with the d ata their experiments generate. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Python for Climate Science: Using Intake to provide easy access to Climate Model data - Charles Turner URL:https://cfp.scipy.org/scipy2025/talk/BJL3U3/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-HK3AAQ@cfp.scipy.org DTSTART;TZID=PST:20250709T112500 DTEND;TZID=PST:20250709T115500 DESCRIPTION:Data manipulation libraries like Polars allow us to analyze and process data much faster than with native Python\, but that’s only true if you know how to use them properly. When the team working on NCEI's Glo bal Summary of the Month first integrated Polars\, they found it was actua lly slower than the original Java version. In this talk\, we'll discuss ho w our team learned how to think about computing problems like spreadsheet programmers\, increasing our products’ processing speed by over 80%. We ’ll share tips for rewriting legacy code to take advantage of parallel p rocessing. We’ll also cover how we created custom\, pre-compiled functio ns with Numba when the business requirements were too complex for native P olars expressions. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Breaking Out of the Loop: Refactoring Legacy Software with Polars - Brodie Vidrine URL:https://cfp.scipy.org/scipy2025/talk/HK3AAQ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-TJNNBD@cfp.scipy.org DTSTART;TZID=PST:20250709T112500 DTEND;TZID=PST:20250709T115500 DESCRIPTION:Image analysis is a central tool in modern biology. Cell and de velopmental biologists generate multidimensional microscopy data\, includi ng imaging of cellular\, subcellular and tissue structures\, in three dime nsions\, over time\, and with multiple molecular markers. Segmentation and tracking of multidimensional microscopy data requires high accuracy acros s many images (e.g. timepoints) and is a labour-intensive part of biologic al image processing pipelines. We present ReSCU-Nets\, recurrent convoluti onal neural networks that use the segmentation results from the previous f rame as a prompt to segment the current frame. We demonstrate that ReSCU-N ets outperform state-of-the-art segmentation models in different tasks on biological multidimensional microscopy sequences. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:ReSCU-Nets: recurrent U-Nets for segmentation of multidimensional m icroscopy data - Rodrigo Fernandez-Gonzalez\, Raymond Hawkins URL:https://cfp.scipy.org/scipy2025/talk/TJNNBD/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LHMFPL@cfp.scipy.org DTSTART;TZID=PST:20250709T112500 DTEND;TZID=PST:20250709T115500 DESCRIPTION:OpenMC is an open source\, community-developed\, Monte Carlo to ol for neutron transport simulations\, featuring a depletion module for fu el burnup calculations in nuclear reactors and a Python API. Depletion cal culations can be expensive as they require solving the neutron transport a nd bateman equations in each timestep to update the neutron flux and mater ial composition\, respectively. Material properties such as temperature an d density govern material cross sections\, which in turn govern reaction r ates. The reaction rates can effect the neutron population. In a scenario where there is no significant change in the material properties or composi tion\, the transport simulation may only need to be run once\; the same cr oss sections are used for the entire depletion calculation. We recently ex tended the depletion module in OpenMC to enable transport-independent depl etion using multigroup cross sections and fluxes. This talk will focus on the technical details of this feature\, its validation\, and briefly touch on areas where the feature has been used. Two recent use cases will be hi ghlighted. The first use case calculates shutdown dose rates for fusion po wer applications\, and the second performs depletion for fission reactor f uel cycle modeling. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Burning fuel for cheap! Transport-independent depletion in OpenMC - Oleksandr Yardas URL:https://cfp.scipy.org/scipy2025/talk/LHMFPL/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-9MUQMM@cfp.scipy.org DTSTART;TZID=PST:20250709T112500 DTEND;TZID=PST:20250709T115500 DESCRIPTION:# GBNet\nGradient Boosting Machines (GBMs) are widely used for their predictive power and interpretability\, while Neural Networks offer flexible architectures but can be opaque. GBNet is a Python package that i ntegrates XGBoost and LightGBM with PyTorch. By leveraging PyTorch’s aut o-differentiation\, GBNet enables novel architectures for GBMs that were p reviously exclusive to pure Neural Networks. The result is a greatly expan ded set of applications for GBMs and an improved ability to interpret expr essive architectures due to the use of GBMs. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:GBNet: Gradient Boosting packages integrated into PyTorch - Michael Horrell URL:https://cfp.scipy.org/scipy2025/talk/9MUQMM/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-3WN9F9@cfp.scipy.org DTSTART;TZID=PST:20250709T121500 DTEND;TZID=PST:20250709T130000 DESCRIPTION:The widespread fascination with AI often fuels a "myth of the a rtificial"\, the belief that scientific and technological progress stems s olely from algorithms and large tech breakthroughs. This talk challenges t hat notion\, arguing that truly responsible and impactful science is funda mentally built upon and sustained by the resilient\, collective intelligen ce of the scientific and research community. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:The Myth of Artificial: Spotlighting Community Intelligence for Res ponsible Science - Dr. Malvika Sharan URL:https://cfp.scipy.org/scipy2025/talk/3WN9F9/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-QWLQWF@cfp.scipy.org DTSTART;TZID=PST:20250709T131500 DTEND;TZID=PST:20250709T134500 DESCRIPTION:This talk walks all Pythonistas through recent CuPy feature dev elopment. Join me and hear my story on how an open-source novice started c ontributing to and helping CuPy over the years grow into a full-fledged\, reliable\, GPU-accelerated array library that covers most of NumPy\, SciPy \, and Numba functionalities. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:CuPy: My Journey toward GPU-Accelerated Computing in Python - Leo F ang URL:https://cfp.scipy.org/scipy2025/talk/QWLQWF/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-P8B77T@cfp.scipy.org DTSTART;TZID=PST:20250709T131500 DTEND;TZID=PST:20250709T134500 DESCRIPTION:Cubed is a framework for distributed processing of large arrays without a cluster. Designed to respect memory constraints at all times\, Cubed can express any NumPy-like array operation as a series of embarrassi ngly-parallel\, bounded-memory steps. By using Zarr as persistent storage between steps\, Cubed can run in a serverless fashion on both a local mach ine and on a range of Cloud platforms. After explaining Cubed’s model\, we will show how Cubed has been integrated with Xarray and demonstrate its performance on various large array geoscience workloads. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Cubed: Scalable array processing with bounded-memory in Python - To m Nicholas\, Tom White URL:https://cfp.scipy.org/scipy2025/talk/P8B77T/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-MHMTLS@cfp.scipy.org DTSTART;TZID=PST:20250709T131500 DTEND;TZID=PST:20250709T134500 DESCRIPTION:Many notable PyData projects including Jupyter Hub\, Matplotlib and JAX follow a versioning scheme called EffVer\, where instead of makin g promises around backward compatibility they communicate the likelihood a nd magnitude of the work required to adopt a new version.\n\nIn this talk we will dive into [EffVer](https://jacobtomlinson.dev/effver/)\, what it i s and what it means for developers and users. We will discuss how to apply EffVer to your own projects and how to depend on projects that use it. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:EffVer: Versioning code by the effort required to upgrade - Jacob T omlinson URL:https://cfp.scipy.org/scipy2025/talk/MHMTLS/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LEHUMF@cfp.scipy.org DTSTART;TZID=PST:20250709T131500 DTEND;TZID=PST:20250709T134500 DESCRIPTION:Training Large Language Models (LLMs) requires processing massi ve-scale datasets efficiently. Traditional CPU-based data pipelines strugg le to keep up with the exponential growth of data\, leading to bottlenecks in model training. In this talk\, we present NeMo Curator\, an accelerate d\, scalable Python-based framework designed to curate high-quality datase ts for LLMs efficiently. Leveraging GPU-accelerated processing with RAPIDS \, NeMo Curator provides modular pipelines for synthetic data generation\, deduplication\, filtering\, classification\, and PII redaction—improvin g data quality and training efficiency. \n\nWe will showcase real-world ex amples demonstrating how multi-node\, multi-GPU processing scales dataset preparation to 100+ TB of data\, achieving up to 7% improvement in LLM dow nstream tasks. Attendees will gain insights into configurable pipelines th at enhance training workflows\, with a focus on reproducibility\, scalabil ity\, and open-source integration within Python's scientific computing eco system. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Unlocking AI Performance with NeMo Curator: Scalable Data Processin g for LLMs - Allison Ding URL:https://cfp.scipy.org/scipy2025/talk/LEHUMF/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-GJRGVU@cfp.scipy.org DTSTART;TZID=PST:20250709T135500 DTEND;TZID=PST:20250709T142500 DESCRIPTION:Large language models (LLMs) enable powerful data-driven applic ations\, but many projects get stuck in “proof-of-concept purgatory” —where flashy demos fail to translate into reliable\, production-ready s oftware. This talk introduces the **LLM software development lifecycle (SD LC)**—a structured approach to moving beyond early-stage prototypes. Usi ng first principles from software engineering\, observability\, and iterat ive evaluation\, we’ll cover common pitfalls\, techniques for structured output extraction\, and methods for improving reliability in real-world d ata applications. Attendees will leave with concrete strategies for integr ating **AI into scientific Python workflows**—ensuring LLMs generate val ue beyond the prototype stage. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Escaping Proof-of-Concept Purgatory: Building Robust LLM-Powered Ap plications - hugo bowne-anderson URL:https://cfp.scipy.org/scipy2025/talk/GJRGVU/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-ELZLHP@cfp.scipy.org DTSTART;TZID=PST:20250709T135500 DTEND;TZID=PST:20250709T142500 DESCRIPTION:Over the past few years\, Discrete Global Grid Systems (DGGS) t hat subdivide the earth into (roughly) equally sized faces have seen a ris e in popularity. However\, their in-memory representation is different fro m traditional projection-based data\, which is either comprised of evenly shaped rectangular grid (aka raster) or discrete geometries (aka vector)\, and thus requires specialized tooling. In particular\, this includes libr aries that can work on the numeric cell ids defined by the specific DGGS.\ n\n`xdggs` is a library that provides a unified interface for `xarray` tha t allows working with and visualizing a variety of DGGS-indexed data sets. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Using Discrete Global Grid Systems in the Pangeo ecosystem - Alexan der Kmoch\, Benoît Bovy\, Jean-Marc Delouis\, Anne Fouilloux\, Justus Mag in\, Tina Odaka URL:https://cfp.scipy.org/scipy2025/talk/ELZLHP/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LQNTGC@cfp.scipy.org DTSTART;TZID=PST:20250709T135500 DTEND;TZID=PST:20250709T142500 DESCRIPTION:For the past decade\, SQL has reigned king of the data transfor mation world\, and tools like dbt have formed a cornerstone of the modern data stack. Until recently\, Python-first alternatives couldn't compete wi th the scale and performance of modern SQL. Now Ibis can provide the same benefits of SQL execution with a flexible Python dataframe API.\n\nIn this talk\, you will learn how Ibis supercharges existing open-source librarie s like Kedro and Pandera and how you can combine these technologies (and a few more) to build and orchestrate scalable data engineering pipelines wi thout sacrificing the comfort (and other advantages) of Python. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Python is all you need: an overview of the composable\, Python-nati ve data stack - Deepyaman Datta URL:https://cfp.scipy.org/scipy2025/talk/LQNTGC/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RECJVV@cfp.scipy.org DTSTART;TZID=PST:20250709T135500 DTEND;TZID=PST:20250709T142500 DESCRIPTION:One of the most important aspects of developing scientific soft ware is distribution for others. The *Scientific Python Development Guide* was developed to provide up-to-date best practices for packaging\, lintin g\, and testing\, along with a versatile template supporting multiple back ends\, and a WebAssembly-powered repo-review tool to check a repository di rectly in the guide. This talk\, with the guide for reference\, will cover key best practices for project setup\, backend selection\, packaging meta data\, GitHub Actions for testing and deployment\, tools for validating co de quality. We will even cover tools for packaging compiled components tha t are simple enough for anyone to use. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Packaging a Scientific Python Project - Henry Fredrick Schreiner II I URL:https://cfp.scipy.org/scipy2025/talk/RECJVV/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-ETWXLC@cfp.scipy.org DTSTART;TZID=PST:20250709T143500 DTEND;TZID=PST:20250709T150500 DESCRIPTION:Tracking and Object-Based Analysis of Clouds (tobac) is a Pytho n package that enables researchers to identify\, track\, and perform objec t-based analyses of phenomena in large atmospheric datasets. Over the past four years\, tobac’s userbase has grown within atmospheric science\, an d the package has transitioned from its original life as a small\, focused package with few maintainers to a larger package with more robust governa nce and structure. In this presentation\, we will discuss the challenges a nd lessons learned during the transition to robust governance structures a nd the future of tobac as we incorporate new techniques for using multiple variables and scales to track the same system. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:tobac: Tracking Atmospheric Phenomena on Multiscale\, Multivariate Diverse Datasets - Sean W. Freeman URL:https://cfp.scipy.org/scipy2025/talk/ETWXLC/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-WM9UFJ@cfp.scipy.org DTSTART;TZID=PST:20250709T143500 DTEND;TZID=PST:20250709T150500 DESCRIPTION:LLMs are powerful\, flexible\, easy-to-use... and often wrong. This is a dangerous combination\, especially for data analysis and scienti fic research\, where correctness and reproducibility are core requirements . Fortunately\, it turns out that by carefully applying LLMs to narrower u se cases\, we can turn them into surprisingly reliable assistants that acc elerate and enhance\, rather than undermine\, scientific work.\n\nThis is not just theory—I’ll showcase working examples of seamlessly integrati ng LLMs into analytic workflows\, helping data scientists build interactiv e\, intelligent applications without needing to be web developers. You’l l see firsthand how keeping LLMs focused lets us leverage their "intellige nce" in a way that’s practical\, rigorous\, and reproducible. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Keeping LLMs in Their Lane: Focused AI for Data Science and Researc h - Joe Cheng URL:https://cfp.scipy.org/scipy2025/talk/WM9UFJ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-NRMNDX@cfp.scipy.org DTSTART;TZID=PST:20250709T143500 DTEND;TZID=PST:20250709T150500 DESCRIPTION:User guides are the piece you often hit right after clicking th e "Learn" or "Get Started" button in a package's documentation. They're re sponsible for onboarding new users\, and providing a learning path through a package. Surprisingly\, while pieces of documentation like the API Refe rence tend to be the same\, the design of user guides tend to differ acros s packages.\n\nIn this talk\, I'll discuss how to design an effective user guide for open source software. I'll explain how the guides for Polars\, DuckDB\, and FastAPI balance working end-to-end like a course\, with being browsable like a reference. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:User guides: engaging new users\, delighting old ones - Michael Cho w URL:https://cfp.scipy.org/scipy2025/talk/NRMNDX/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-PBLESZ@cfp.scipy.org DTSTART;TZID=PST:20250709T143500 DTEND;TZID=PST:20250709T150500 DESCRIPTION:Block-based programming divides inputs into local arrays that a re processed concurrently by groups of threads. Users write sequential arr ay-centric code\, and the framework handles parallelization\, synchronizat ion\, and data movement behind the scenes. This approach aligns well with SciPy's array-centric ethos and has roots in older HPC libraries\, such as [NWChem’s TCE](https://nwchemgit.github.io/TCE.html)\, [BLIS](https://g ithub.com/flame/blis)\, and [ATLAS](https://math-atlas.sourceforge.net/). \n\nIn recent years\, many block-based Python programming models for GPUs have emerged\, like [Triton](https://openai.com/index/triton/)\, [JAX/Pall as](https://docs.jax.dev/en/latest/pallas/index.html)\, and [Warp](https:/ /nvidia.github.io/warp/modules/tiles.html)\, aiming to make parallelism mo re accessible for scientists and increase portability.\n\n**In this talk\, we'll present [cuTile and Tile IR](https://x.com/blelbach/status/19021137 67066103949)\, a new Pythonic tile-based programming model and compiler re cently announced by NVIDIA.** We'll explore cuTile examples from a variety of domains\, including a new [LLAMA3](https://github.com/meta-llama/llama 3)-based reference app and a port of [miniWeather](https://github.com/mrno rman/miniWeather). You'll learn the best practices for writing and debuggi ng block-based Python GPU code\, gain insight into how such code performs\ , and learn how it differs from traditional SIMT programming. \n\nBy the e nd of the session\, you'll understand how block-based GPU programming enab les more intuitive\, portable\, and efficient development of high-performa nce\, data-parallel Python applications for HPC\, data science\, and machi ne learning. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:cuTile\, the New/Old Kid on the Block: Python Programming Models fo r GPUs - Bryce Adelstein Lelbach URL:https://cfp.scipy.org/scipy2025/talk/PBLESZ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-XMC8KU@cfp.scipy.org DTSTART;TZID=PST:20250709T152500 DTEND;TZID=PST:20250709T155500 DESCRIPTION:Generative Artificial Intelligence (AI) is reshaping engineerin g education by\noffering students new ways to engage with complex concepts and content. Ethical\nconcerns including bias\, intellectual property\, a nd plagiarism make Generative AI\na controversial educational tool. Overre liance on AI may also lead to academic\nintegrity issues\, necessitating c lear student codes of conduct that define acceptable\nuse. As educators we should carefully design learning objectives to align with\ntransferrable career skills in our fields. By practicing backward design with a\nfocus o n career-readiness skills\, we can incorporate useful prompt engineering\, \nrapid prototyping\, and critical reasoning skills that incorporate gener ative AI.\nEngineering students want to develop essential career skills su ch as critical\nthinking\, communication\, and technology. This talk will focus on case studies for\nusing generative AI and rapid prototyping for s cientific computing in engineering\ncourses for physics\, programming\, an d technical writing. These courses include\nassignments and reading exampl es using NumPy\, SciPy\, Pandas\, etc. in Jupyter\nnotebooks. Embracing ge nerative AI tools has helped students compare\, evaluate\,\nand discuss wo rk that was inaccessible before generative AI. This talk explores\nstrateg ies for using AI in engineering education while accomplishing learning\nob jectives and giving students opportunities to practice career readiness sk ills. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Generative AI in Engineering Education: A Tool for Learning\, Not a Replacement for Skills - Ryan C Cooper URL:https://cfp.scipy.org/scipy2025/talk/XMC8KU/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-8WQQPV@cfp.scipy.org DTSTART;TZID=PST:20250709T152500 DTEND;TZID=PST:20250709T155500 DESCRIPTION:Scaling artificial intelligence (AI) and machine learning (ML) workflows on high-performance computing (HPC) systems presents unique chal lenges\, particularly as models become more complex and data-intensive. Th is study explores strategies to optimize AI/ML workflows for enhanced perf ormance and resource utilization on HPC platforms.​\n\nWe investigate ad vanced parallelization techniques\, such as Data Parallelism (DP)\, Distri buted Data Parallel (DDP)\, and Fully Sharded Data Parallel (FSDP). Implem enting memory-efficient strategies\, including mixed precision training an d activation checkpointing\, significantly reduces memory consumption with out compromising model accuracy. Additionally\, we examine various communi cation backends( i.e. NCCL\, MPI\, and Gloo) to enhance inter-GPU and inte r-node communication efficiency. Special attention is given to the complex ities of implementing these backends in HPC environments\, providing solut ions for proper configuration and execution.​\n\nOur findings demonstrat e that these optimizations enable stable and scalable AI/ML model training and inference\, achieving substantial improvements in training times and resource efficiency. This presentation will detail the technical challenge s encountered and the solutions developed\, offering insights into effecti vely scaling AI/ML workflows on HPC systems.​ DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Scaling AI/ML Workflows on HPC for Geoscientific Applications. - Ne gin Sobhani URL:https://cfp.scipy.org/scipy2025/talk/8WQQPV/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-DSYLVL@cfp.scipy.org DTSTART;TZID=PST:20250709T152500 DTEND;TZID=PST:20250709T155500 DESCRIPTION:X-ray ptychographic imaging is becoming an indispensable tool f or visualizing matter at nanoscale\, driving innovation across many fields \, including functional materials\, electronics\, life sciences\, etc. Thi s imaging mode is particularly attractive thanks to its ability to generat e high-resolution view of an extended object without using a lens with hig h numerical aperture. The technique relies on advanced mathematical algori thms to retrieve the missing phase information that is not directly record ed by a physical detector\, therefore computation intensive. Advances in a ccelerator\, optics\, and detector technologies have greatly increased dat a generate rate\, imposing a big challenge on efficient execution of recon struction process to support decision-making in an experiment. Here\, we d emonstrate how efficient GPU-based reconstruction algorithms\, deployed at the edge\, enable real-time feedback during high-speed continuous data ac quisition increasing the speed and efficiency of the experiments. The deve lopments further pave the way for AI-augmented autonomous microscopic expe rimentation performed at machine speeds. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Edge processing of X-ray ptychography: enabling real-time feedback for high-speed data acquisition - Seher Karakuzu\, Denis Leshchev URL:https://cfp.scipy.org/scipy2025/talk/DSYLVL/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LS3LFX@cfp.scipy.org DTSTART;TZID=PST:20250709T152500 DTEND;TZID=PST:20250709T155500 DESCRIPTION:At CERN (European Organization for Nuclear Research)\, machine learning models are developed and deployed for various applications\, incl uding data analysis\, event reconstruction\, and classification. These mod els must not only be highly sophisticated but also optimized for efficient inference. A critical application is in Triggers- systems designed to ide ntify and select interesting events from an immense stream of experimental data. Experiments like ATLAS and CMS generate data at rates of approximat ely 100 TB/s\, requiring Triggers to rapidly filter out irrelevant events. This talk will explore the challenges of deploying machine learning in su ch high-throughput environments and discuss solutions to enhance their per formance and reliability. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Challenges and Implementations for ML Inference in High-energy Phys ics - Sanjiban Sengupta URL:https://cfp.scipy.org/scipy2025/talk/LS3LFX/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-XWYKHA@cfp.scipy.org DTSTART;TZID=PST:20250709T160500 DTEND;TZID=PST:20250709T163500 DESCRIPTION:This talk presents a candid reflection on integrating generativ e AI into an Engineering Computations course\, revealing unexpected challe nges despite best intentions. Students quickly developed patterns of using AI as a shortcut rather than a learning companion\, leading to decreased attendance and an "illusion of competence." I'll discuss the disconnect be tween instructor expectations and student behavior\, analyze how tradition al assessment structures reinforced counterproductive AI usage\, and share strategies for guiding students toward using AI as a co-pilot rather than a substitute for critical thinking while maintaining academic integrity. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Embracing GenAI in Engineering Education: Lessons from the Trenches - Lorena Barba URL:https://cfp.scipy.org/scipy2025/talk/XWYKHA/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-8SRJ3V@cfp.scipy.org DTSTART;TZID=PST:20250709T160500 DTEND;TZID=PST:20250709T163500 DESCRIPTION:Explainable AI (XAI) emerged to clarify the decision-making of complex deep learning models\, but standard XAI methods are often uninform ative on Earth system models due to their high-dimensional and physically constrained nature. We introduce “physical XAI\,” which adapts XAI tec hniques to maintain physical realism and handle autocorrelated data effect ively. Our approach includes physically consistent perturbations\, analysi s of uncertainty\, and the use of variance-based global sensitivity tools. Furthermore\, we expand the definition of “physical XAI” to include m eaningful interactive data analysis. We demonstrate these methods on two E arth system models: a data-driven global weather model and a winter precip itation type model to show how we can gain more physically meaningful insi ghts. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Physical XAI - Going Beyond Traditional XAI Methods in Earth System Science - Charlie Becker URL:https://cfp.scipy.org/scipy2025/talk/8SRJ3V/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-ESGGCR@cfp.scipy.org DTSTART;TZID=PST:20250709T160500 DTEND;TZID=PST:20250709T163500 DESCRIPTION:Many scientists rely on NumPy for its simplicity and strong CPU performance\, but scaling beyond a single node is challenging. The resear chers at SLAC need to process massive datasets under tight beam time const raints\, often needing to modify code on the fly. This is where cuPyNumeri c comes in—a drop-in replacement for NumPy that distributes work across CPUs and GPUs. With its familiar NumPy interface\, cuPyNumeric makes it ea sy to scale computations without rewriting code\, helping scientists focus on their research instead of debugging. It’s a great example of how the SciPy ecosystem enables cutting-edge science. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Scaling NumPy for Large-Scale Science: The cuPyNumeric Approach - I rina Demeshko\, Quynh L. Nguyen URL:https://cfp.scipy.org/scipy2025/talk/ESGGCR/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RM3UHE@cfp.scipy.org DTSTART;TZID=PST:20250709T160500 DTEND;TZID=PST:20250709T163500 DESCRIPTION:Computational needs in high energy physics applications are inc reasingly met by utilizing GPUs as hardware accelerators\, but achieving t he highest throughput requires directly reading data into GPU memory. This has yet to be achieved for HEP’s standard domain specific “ROOT” fi le formats. Using KvikIO’s python bindings to CuFile and NvComp\, KvikUp root is a prototype package to support the reading of ROOT file formats by the GPU. On GPUDirect storage (GDS) enabled systems\, data bypasses the C PU and is loaded directly from storage to the GPU. We will discuss the met hodology we developed to read ROOT files into GPUs via RDMA. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:KvikUproot - Reading and Deserializing High Energy Physics Data wit h KvikIO and CuPy - Frank Strug URL:https://cfp.scipy.org/scipy2025/talk/RM3UHE/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-7KXVCV@cfp.scipy.org DTSTART;TZID=PST:20250709T170000 DTEND;TZID=PST:20250709T180000 DESCRIPTION:Lightning talks are 5-minute talks on any topic of interest for the SciPy community. We encourage spontaneous and prepared talks from eve ryone\, but we can’t guarantee spots. Sign ups are at the NumFOCUS booth during the conference. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Lightning Talks - URL:https://cfp.scipy.org/scipy2025/talk/7KXVCV/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-EWLSUX@cfp.scipy.org DTSTART;TZID=PST:20250709T180000 DTEND;TZID=PST:20250709T180000 DESCRIPTION:The virtual poster session takes place 6:00–7:00 p.m. in the SciPy 2025 Gather space. Meet with the poster authors to ask questions and learn about the posters that will be on display in the poster hall inside our virtual space. Virtual and in-person ticket holders are all welcome t o attend in Gather! DTSTAMP:20260417T070310Z LOCATION:Virtual SUMMARY:(Exclusively in Gather) Virtual Poster Session - URL:https://cfp.scipy.org/scipy2025/talk/EWLSUX/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-WDUNLN@cfp.scipy.org DTSTART;TZID=PST:20250709T180000 DTEND;TZID=PST:20250709T190000 DESCRIPTION:The Poster session will be in the Ballroom from 6:00-7:00pm. Me et with the poster authors to ask questions and learn about the posters th at will be on display throughout the main conference. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:SciPy 2025 Poster Session - URL:https://cfp.scipy.org/scipy2025/talk/WDUNLN/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-GLXVW8@cfp.scipy.org DTSTART;TZID=PST:20250709T190000 DTEND;TZID=PST:20250709T213000 DESCRIPTION:Join us for an evening of connection and creativity at the Muse um of Glass\, just a short walk from the conference venue. Explore stunnin g glass art exhibits\, enjoy refreshments\, and connect with fellow SciPy attendees in a unique and inspiring setting. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Attendee Social at the Museum of Glass - hosted by NVIDIA - URL:https://cfp.scipy.org/scipy2025/talk/GLXVW8/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RBD9KX@cfp.scipy.org DTSTART;TZID=PST:20250710T091500 DTEND;TZID=PST:20250710T100000 DESCRIPTION:This keynote will trace the personal journey of NumPy's develop ment and the evolution of the SciPy community from 2001 to the present. Dr awing on over two decades of involvement\, I’ll reflect on how a small g roup of enthusiastic contributors grew into a vibrant\, global ecosystem t hat now forms the foundation of scientific computing in Python. Through st ories\, milestones\, and community moments\, we’ll explore the challenge s\, breakthroughs\, and collaborative spirit that shaped both NumPy and th e SciPy conventions over the years. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:My Dinner with Numeric\, Numpy\, and Scipy: A Retrospective from 20 01 to 2025 with Comments and Anecdotes - Charles R Harris URL:https://cfp.scipy.org/scipy2025/talk/RBD9KX/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-AARA39@cfp.scipy.org DTSTART;TZID=PST:20250710T104500 DTEND;TZID=PST:20250710T111500 DESCRIPTION:Xarray has enormous potential as a data model and toolkit for l abeled N-D arrays in biology. Originally developed within the geosciences community\, it is seeing increased usage in biology\, with applications ra nging from genomics to image analysis and beyond. However\, it has not yet been widely adopted. This presentation will investigate what the blockers have been to wider adoption\, showcase the power of Xarray in biology thr ough existing use cases\, and present a roadmap for the future of Xarray i n biological workflows through recent and upcoming improvements in Xarray. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Xarray across biology. Where are we and where are we going? - Ian H unt-Isaak URL:https://cfp.scipy.org/scipy2025/talk/AARA39/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RHKQKD@cfp.scipy.org DTSTART;TZID=PST:20250710T104500 DTEND;TZID=PST:20250710T111500 DESCRIPTION:Rydberg atoms offer unique quantum properties that enable radio -frequency sensing capabilities distinct from any classical analogue\; how ever\, large parameter spaces and complex configurations make understandin g and designing these quantum experiments challenging. Current solutions a re often developed as in-house\, closed-sourced software simulating a narr ow range of problems. We present RydIQule\, an open-source package leverag ing tools of computational python in novel ways to model the behavior of t hese systems generally. We describe RydIQule’s approach to representing quantum systems using computational graphs and leveraging numpy broadcasti ng to define complete experiments. In addition to discussing the computati onal challenges RydIQule helps overcome\, we outline how collaboration bet ween physics and computational research backgrounds has led to this impact ful tool. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:RydIQule: A Package for Modelling Quantum Sensors - Ben Miller URL:https://cfp.scipy.org/scipy2025/talk/RHKQKD/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-A3UYEC@cfp.scipy.org DTSTART;TZID=PST:20250710T104500 DTEND;TZID=PST:20250710T111500 DESCRIPTION:This talk explores various methods to accelerate traditional ma chine learning pipelines using scikit-learn\, UMAP\, and HDBSCAN on GPUs. We will contrast the experimental Array API Standard support layer in scik it-learn with the cuML library from the NVIDIA RAPIDS Data Science stack\, including its zero-code change acceleration capability. ML and data scien ce practitioners will learn how to seamlessly accelerate machine learning workflows\, highlight performance benefits\, and receive practical guidanc e for different problem types and sizes. Insights into minimizing cost and runtime by effectively mixing hardware for various tasks\, as well as the current implementation status and future plans for these acceleration met hods\, will be provided. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:GPUs & ML – Beyond Deep Learning - Simon Adorf URL:https://cfp.scipy.org/scipy2025/talk/A3UYEC/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-UDTLL7@cfp.scipy.org DTSTART;TZID=PST:20250710T104500 DTEND;TZID=PST:20250710T111500 DESCRIPTION:Designing tomorrow's materials requires understanding how atoms behave – a challenge that's both fascinating and incredibly complex. Wh ile machine learning offers exciting speedups in materials simulation\, it often falls short\, missing vital electronic structure information needed to connect theory with experimental results. This work introduces a power ful solution: Density Functional Tight Binding (DFTB)\, which\, combined w ith the versatile tools of Scientific Python\, allows us to understand the electronic behavior of materials while maintaining computational efficien cy. In this talk\, I will present our findings demonstrating how DFTB\, co upled with readily available Python packages\, allows for direct compariso n between theoretical predictions and experimental data\, such as XPS meas urements. I will also showcase our publicly available repository\, contai ning DFTB parameters for a wide range of materials\, making this powerful approach accessible to the broader research community. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Can Scientific Python Tools Unlock the Secrets of Materials? The El ectrons That Machine-Learning Can't Handle - Filippo Balzaretti URL:https://cfp.scipy.org/scipy2025/talk/UDTLL7/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-RXCE3F@cfp.scipy.org DTSTART;TZID=PST:20250710T112500 DTEND;TZID=PST:20250710T115500 DESCRIPTION:Synthetic aviation fuels (SAFs) offer a pathway to improving ef ficiency\, but high cost and volume requirements hinder property testing a nd increase risk of developing low-performing fuels. To promote productive SAF research\, we used Fourier Transform Infrared (FTIR) spectra to train accurate\, interpretable fuel property models. In this presentation\, we will discuss how we leveraged standard Python libraries – NumPy\, pandas \, and scikit-learn – and Non-negative Matrix Factorization to decompose FTIR spectra and develop predictive models. Specifically\, we will review the pipeline developed for preprocessing FTIR data\, the ensemble models used for property prediction\, and how the features correlate with physico chemical properties. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Advanced Machine Learning Techniques for Predicting Properties of S ynthetic Aviation Fuels using Python - Ana Comesana URL:https://cfp.scipy.org/scipy2025/talk/RXCE3F/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-CWJ7XR@cfp.scipy.org DTSTART;TZID=PST:20250710T112500 DTEND;TZID=PST:20250710T115500 DESCRIPTION:The “napari-activelearning” plugin provides a framework to fine tune deep learning models for large-scale bioimage analysis\, such as digital pathology Whole Slide Images (WSI). This plugin was developed wit h the motivation of easing the integration of deep learning tools into bio image analysis workflows. This plugin implements the concept of Active Lea rning for reducing the time spent on labeling samples when fine tuning mod els. Because this plugin is integrated into Napari and leverages the use o f next generation file formats (Zarr)\, it is suitable for fine tuning dee p learning models on large-scale images with little image preparation. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:An Active Learning plugin in Napari to fine tune models for large-s cale bioimage analysis - Fernando Cervantes Sanchez URL:https://cfp.scipy.org/scipy2025/talk/CWJ7XR/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-7XRAQD@cfp.scipy.org DTSTART;TZID=PST:20250710T112500 DTEND;TZID=PST:20250710T115500 DESCRIPTION:AI assistants are evolving from simple Q&A bots to intelligent\ , multimodal\, multilingual\, and agentic systems capable of reasoning\, r etrieving\, and autonomously acting. In this talk\, we’ll showcase how t o build a voice-enabled\, multilingual\, multimodal RAG (Retrieval-Augment ed Generation) assistant using Gradio\, OpenAI’s Whisper\, LangChain\, L angGraph\, and FAISS. Our assistant will not only process voice and text i nputs in multiple languages but also intelligently retrieve information fr om structured and unstructured data. We’ll demonstrate this with a fligh t search use case—leveraging a flight database for retrieval and\, when necessary\, autonomously searching external sources using LangGraph. You w ill gain practical insights into building scalable\, adaptive AI assistant s that move beyond static chatbots to autonomous agents that interact dyna mically with users and the web. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Polyglot RAG: Building a Multimodal\, Multilingual\, and Agentic AI Assistant - Axel Sirota URL:https://cfp.scipy.org/scipy2025/talk/7XRAQD/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-U9GYS3@cfp.scipy.org DTSTART;TZID=PST:20250710T112500 DTEND;TZID=PST:20250710T115500 DESCRIPTION:This talk presents zfit with the newest improvements\, a genera l purpose distribution fitting library for complicated model building beyo nd fitting a normal distribution. The talk will cover all aspects of fitti ng with a focus on the strong model building part in zfit\; composable dis tributions with sums\, products and more\, build and mix binned and unbinn ed\, analytic and templated functions in multiple dimensions. This include s the creation of arbitrary\, custom distributions with minimal effort tha t fulfils everyones need.\nThanks to the numpy-like backend used by Tensor Flow\, zfit is highly performant by using JIT compiled code on CPUs and ev en GPUs\, a showcase for scientific computing faster than numpy. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:zfit: scalable pythonic likelihood fitting - Jonas Eschle URL:https://cfp.scipy.org/scipy2025/talk/U9GYS3/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-7GHVZ3@cfp.scipy.org DTSTART;TZID=PST:20250710T121500 DTEND;TZID=PST:20250710T131500 DESCRIPTION:NVIDIA’s CUDA platform has long been the backbone of high-per formance GPU computing\, but its power has historically been gated behind C and C++ expertise. With the recent introduction of native Python support \, CUDA is more accessible to the programming language you know and love\, ushering in a new era for scientific computing\, data science\, and AI de velopment. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Python at the Speed of Light: Accelerating Science with CUDA Python - Christopher Lamb\, VP of Software for Compute Platforms at NVIDIA URL:https://cfp.scipy.org/scipy2025/talk/7GHVZ3/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-3ZHYMH@cfp.scipy.org DTSTART;TZID=PST:20250710T131500 DTEND;TZID=PST:20250710T141000 DESCRIPTION:Collaborating on code and software is essential to open science —but it’s not always easy. Join this BoF for an interactive discussion on the real-world challenges of open source collaboration. We’ll explor e common hurdles like Python packaging\, contributing to existing codebase s\, and emerging issues around LLM-assisted development and AI-generated s oftware contributions.\n\nWe’ll kick off with a brief overview of pyOpen Sci—an inclusive community of Pythonistas\, from novices to experts—wo rking to make it easier to create\, find\, share\, and contribute to reusa ble code. We’ll then facilitate small-group discussions and use an inter active Mentimeter survey to help you share your experiences and ideas.\n\n Your feedback will directly shape pyOpenSci’s priorities for the coming year\, as we build new programs and resources to support your work in the Python scientific ecosystem. Whether you’re just starting out or a seaso ned developer\, you’ll leave with clear ways to get involved and make an impact on the broader Python ecosystem in service of advancing scientific discovery. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Open Code\, Open Science: What’s Getting in Your Way? - Jeremiah Paige\, Avik Basu\, Tetsuo Koyama\, Leah Wasser\, Inessa Pawson URL:https://cfp.scipy.org/scipy2025/talk/3ZHYMH/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-TQMURM@cfp.scipy.org DTSTART;TZID=PST:20250710T131500 DTEND;TZID=PST:20250710T141000 DESCRIPTION:Conferences serve as a way to connect groups of humans around c ommon topics of interest. In the open source community\, they have played a critical role in knowledge sharing\, advancing technology\, and fosterin g a sense of community. This is especially true for the global Python comm unity. Times are changing\, the political climate both in the US and abroa d has drastically shifted making gathering in the real world much more com plex. Advancements in technologies have changed the calculus on what is co nsidered quality participation. Join us in this BoF to discuss these chall enges and how we can continue to come together as a community. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Organizing Conferences in These Times - Julie Hollek URL:https://cfp.scipy.org/scipy2025/talk/TQMURM/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-EWQ9B8@cfp.scipy.org DTSTART;TZID=PST:20250710T131500 DTEND;TZID=PST:20250710T141500 DESCRIPTION:You'll be randomly paired with another conference attendee for a 5-minute chat. Non-cheesy icebreakers will be provided. Virtual and in-p erson attendees welcome!\n\nZoom link: https://numfocus-org.zoom.us/j/8747 5693685?pwd=XehxS6v7cYI63UsS2h9nBsrGzAhcDs.1 \n2025-07-10 13:15 until 2025 -07-10 14:15 DTSTAMP:20260417T070310Z LOCATION:Virtual SUMMARY:(Exclusively on Zoom) Virtual Speed Networking - URL:https://cfp.scipy.org/scipy2025/talk/EWQ9B8/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-EHA3PC@cfp.scipy.org DTSTART;TZID=PST:20250710T131500 DTEND;TZID=PST:20250710T141000 DESCRIPTION:Generative AI has rapidly changed the landscape of computing an d data education. Many learners are utilizing generative AI to assist in l earning\, so what should educators do to address the opportunities\, risks \, and potential for their use? The goal of this open discussion session i s to bring together community members to unravel these pressing questions in order to not only improve learning outcomes in a variety of diverse con texts: not only students learning in a classroom setting\, but also ed-tec h or generative AI designers developing new user experiences that aim to i mprove human capacities\, and even scientists interested in learning best practices for communicating results to stakeholders or creating learning m aterials for colleagues. The open discussion will include ample opportunit y for community members to network with each other and build connections a fter the conference. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Generative AI in Education - Kevin Lin URL:https://cfp.scipy.org/scipy2025/talk/EHA3PC/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-Z8P8JH@cfp.scipy.org DTSTART;TZID=PST:20250710T142000 DTEND;TZID=PST:20250710T145000 DESCRIPTION:Women remain critically underrepresented in data science and Py thon communities\, comprising only 15–22% of professionals globally and less than 3% of contributors to Python open-source projects. This disparit y not only limits diversity but also represents a missed opportunity for i nnovation and community growth. This talk explores actionable strategies t o address these gaps\, drawing from my leadership in Women in AI at IBM\, TechWomen mentorship\, and initiatives with NumFOCUS. Attendees will gain insights and practical steps to create inclusive environments\, foster div erse collaboration\, and ensure the scientific Python community thrives by unlocking its full potential. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Unlocking the Missing 78%: Inclusive Communities for the Future of Scientific Python - Noor Aftab URL:https://cfp.scipy.org/scipy2025/talk/Z8P8JH/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-XNSXMY@cfp.scipy.org DTSTART;TZID=PST:20250710T142000 DTEND;TZID=PST:20250710T145000 DESCRIPTION:The rapidly evolving Python ecosystem presents increasing chall enges for adapting code using traditional methods. Developers frequently n eed to rewrite applications to leverage new libraries\, hardware architect ures\, and optimization techniques. To address this challenge\, the Numba team is developing a superoptimizing compiler built on equality saturation -based term rewriting. This innovative approach enables domain experts to express and share optimizations without requiring extensive compiler exper tise. This talk explores how Numba v2 enables sophisticated optimizations —from floating-point approximation and automatic GPU acceleration to ene rgy-efficient multiplication for deep learning models—all through the fa miliar NumPy API. Join us to discover how Numba v2 is bringing superoptimi zation capabilities to the Python ecosystem. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Numba v2: Towards a SuperOptimizing Python Compiler - Siu Kwan Lam URL:https://cfp.scipy.org/scipy2025/talk/XNSXMY/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-VWB7YP@cfp.scipy.org DTSTART;TZID=PST:20250710T142000 DTEND;TZID=PST:20250710T145000 DESCRIPTION:This track highlights the fantastic scientific applications tha t the\nSciPy community creates with the tools we collectively make. Talk\n proposals to this track should be stories of how using the Scientific\nPyt hon ecosystem the speakers were able to overcome challenges\, create\nnew collaborations\, reduce the time to scientific insight\, and share\ntheir results in ways not previously possible. Proposals should focus\non novel applications and problems\, and be of broad interest to the\nconference\, but should not shy away from explaining the scientific\nnuances that make the story in the proposal exciting. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Probing the Hidden World of Battery Chemistry With X-rays - Mark Wo lfman URL:https://cfp.scipy.org/scipy2025/talk/VWB7YP/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-NYPUVH@cfp.scipy.org DTSTART;TZID=PST:20250710T142000 DTEND;TZID=PST:20250710T145000 DESCRIPTION:Jupyter Book allows researchers and educators to create books a nd knowledge bases that are reusable\, reproducible\, and interactive. Jup yter Book 2 has been rebuilt on a new document engine that prioritizes ext ensibility\, machine readability and flexible deployment\, allowing us to create and share interactive computational content in new ways. In this ta lk\, we will introduce Jupyter Book 2.0\, demonstrate its game changing fe atures\, and showcase real-world examples like *The Turing Way*\, *QuantEc on* and *Project Pythia*. We'll conclude with a live demo\, taking a folde r of notebooks and markdown files and turning them into a deployable\, fea ture-rich website. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Jupyter Book 2.0 – A Next-Generation tool for sharing for Computa tional Content - Steve Purves\, Rowan Cockett\, Franklin Koch\, Angus Holl ands\, Chris Holdgraf URL:https://cfp.scipy.org/scipy2025/talk/NYPUVH/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-BQMZFM@cfp.scipy.org DTSTART;TZID=PST:20250710T150000 DTEND;TZID=PST:20250710T153000 DESCRIPTION:mybinder.org has served millions of scientific python users for 8 years now! It is an experiment in running open source *infrastructure* as a public good. Sustainability challenges faced by open source *software * production are magnified here - we need people time to manage the infras tructure\, pay for computational infrastructure required to run the servic e\, operate it reliably by responding to outages in a timely fashion\, and fight off abuse from malicious actors. This talk covers the lessons learn t over the years\, and new community oriented experiments to better sustai nability\, functionality & reliability that we are trying out now. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Towards a more sustainable and reliable mybinder.org - Yuvi URL:https://cfp.scipy.org/scipy2025/talk/BQMZFM/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-S77TUP@cfp.scipy.org DTSTART;TZID=PST:20250710T150000 DTEND;TZID=PST:20250710T153000 DESCRIPTION:In today’s world of ever-growing data and AI\, learning about GPUs has become an essential part of software carpentry\, professional de velopment and the education curriculum. However\, teaching with GPUs can b e challenging\, from resource accessibility to managing dependencies and v arying knowledge levels.\n\nDuring this talk we will address these issues by offering practical strategies to promote active learning with GPUs and share our experiences from running numerous Python conference tutorials th at leveraged GPUs. Attendees will learn different options to how to provid e GPU access\, tailor content for different expertise levels\, and simplif y package management when possible.\n\nIf you are an educator\, researcher \, and/or developer who is interested in teaching or learning about GPU co mputing with Python\, this talk will give you the confidence to teach topi cs that require GPU acceleration and quickly get your audience up and runn ing. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Teaching Python with GPUs: Empowering educators to share knowledge that uses GPUs - Jacob Tomlinson\, Naty Clementi URL:https://cfp.scipy.org/scipy2025/talk/S77TUP/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-NEGTPD@cfp.scipy.org DTSTART;TZID=PST:20250710T150000 DTEND;TZID=PST:20250710T153000 DESCRIPTION:Today’s quantum computers are far noisier than their classica l counterparts. Unlike traditional computing errors\, quantum noise is mor e complex\, arising from decoherence\, crosstalk\, and gate imperfections that corrupt quantum states. Error mitigation has become a rapidly evolvin g field\, offering ways to address these errors on existing devices. New t echniques emerge regularly\, requiring flexible tools for implementation a nd testing. This talk explores the challenges of mitigating noise and how researchers and engineers use Python to iterate quickly while maintaining reliable and reproducible workflows. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Noise-Resilient Quantum Computing with Python - nate stemen URL:https://cfp.scipy.org/scipy2025/talk/NEGTPD/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-GJACX9@cfp.scipy.org DTSTART;TZID=PST:20250710T150000 DTEND;TZID=PST:20250710T153000 DESCRIPTION:Reproducibility is a major underpinning of the scientific metho d. In scientific computing\, this also includes the ability to reproduce y our dependencies. Yet\, in 2025 this still remains a challenging topic. \n \nPixi is a modern package manager built on the Conda ecosystem. It integr ates very well with all existing packages on conda-forge. Pixi makes packa ge management reproducible\, fast and painless – so that scientists can go back to coding instead of dealing with “dependency hell”. Pixi impr oves the mix Conda and PyPI package management by integrating with `uv` by astral.sh and streamlines automation with a cross-platform task runner. T hese features combined with a powerful lockfile make creating reproducible projects trivial.\n\n\nThis talk is for people who are interested in new\ , fast ways to set up their software (dev) environments on different syste ms – think your coworker's computer\, CI\, containers\, and more. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Reproducible Science Made Easy: Package Management with Pixi - Wolf Vollprecht\, Ruben Arts URL:https://cfp.scipy.org/scipy2025/talk/GJACX9/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-KAPJYZ@cfp.scipy.org DTSTART;TZID=PST:20250710T155000 DTEND;TZID=PST:20250710T162000 DESCRIPTION:Camera traps are an essential tool for wildlife research. Zamba is an open source Python package that leverages machine learning and comp uter vision to automate time-intensive processing tasks for wildlife camer a trap data. This talk will dive into Zamba's capabilities and key factors that influenced its design and development. Topics will include the impor tance of code-free custom model training\, Zamba’s origins in an open ma chine learning competition\, and the technical challenges of processing vi deo data. Attendees will walk away with a better understanding of how mach ine learning and Python tools can support conservation efforts. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Zamba: Computer vision for wildlife conservation - Emily Dorne\, Ja y Qi URL:https://cfp.scipy.org/scipy2025/talk/KAPJYZ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-JBNR9A@cfp.scipy.org DTSTART;TZID=PST:20250710T155000 DTEND;TZID=PST:20250710T162000 DESCRIPTION:The best way to distribute large scientific datasets is via the Cloud\, in Cloud-Optimized formats. But often this data is stuck in archi val pre-Cloud file formats such as netCDF.\n\nVirtualiZarr makes it easy t o create "Virtual" Zarr datacubes\, allowing performant access to huge arc hival datasets as if it were in the Cloud-Optimized Zarr format\, without duplicating any of the original data.\n\nWe will demonstrate using Virtual iZarr to generate references to archival files\, combine them into one arr ay datacube using xarray-like syntax\, commit them to Icechunk\, and read the data back with zarr-python v3. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:VirtualiZarr and Icechunk: How to build a cloud-optimised datacube of archival files in 3 lines of xarray - Tom Nicholas URL:https://cfp.scipy.org/scipy2025/talk/JBNR9A/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-YKCR8S@cfp.scipy.org DTSTART;TZID=PST:20250710T155000 DTEND;TZID=PST:20250710T162000 DESCRIPTION:The Universe isn't always so quiet: neutron stars\, fast radio bursts\, and potentially alien civilizations emit bursts of electromagneti c energy - radio transients - into the unknown. In some cases\, these emis sions\, like with pulsars\, are constant and periodic\; but in others\, li ke with fast radio bursts\, they're short in duration and infrequent. Clas sical detection surveys typically rely on dedispersion techniques and huma n-crafted signal processing filters to remove noise and highlight a signal of interest. But what if we're missing something?\n\nIn this talk we will introduce a workflow to avoid classical processing all together. By feedi ng RF samples directly from the telescope's digitizers into GPU computing\ , we can train an AI model to serve as a detector -- not only enabling rea l time performance\, but also making decisions directly on raw spectrogram data\, eliminating the need for classical processing. We will demonstrate how each step of the pipeline works - from AI model training and data cur ation to real-time inferencing at scale. Our hope is that this new sensor processing architecture can simplify development\, democratize science\, a nd process increasingly large amounts of data in real time. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:AI as a Detector: Lessons in Real Time Pulsar Discovery - Adam Thom pson\, Luigi Cruz URL:https://cfp.scipy.org/scipy2025/talk/YKCR8S/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-B3PXML@cfp.scipy.org DTSTART;TZID=PST:20250710T155000 DTEND;TZID=PST:20250710T162000 DESCRIPTION:The Issaquah Robotics Society (IRS) has been teaching Python an d data analysis to high school students since 2016. Our presentation will summarize what we’ve learned from nine years of combining Python\, compe titive robotics\, and high school students with no prior programming exper ience. We’ll focus on the importance of keeping it fun\, learning the to ols\, and how to provide useful feedback without making learning Python fe el like just another class. We’ll also explain how Python helps us win r obotics competitions. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Keeping Python Fun: Using Robotics Competitions to Teach Data Analy sis and Application Development - Stacy Irwin URL:https://cfp.scipy.org/scipy2025/talk/B3PXML/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-EMLLYF@cfp.scipy.org DTSTART;TZID=PST:20250710T163000 DTEND;TZID=PST:20250710T170000 DESCRIPTION:In Python\, data analytics users often prioritize convenience\, flexibility\, and familiarity over pure performance. The cuDF DataFrame l ibrary provides a pandas-like experience with from 10x up to 50x performan ce improvements\, but subtle differences prevent it from being a true drop -in replacement for many users. This talk will showcase the evolution of t his library to provide zero-code change experiences\, first for pandas use rs and now for Polars. We will provide examples of this usage and a high l evel overview of how users can make use of these today. We will then delve into the details of how GPU acceleration is implemented differently in pa ndas and Polars\, along with a deep dive into some of the different techni cal challenges encountered for each. This talk will have something for bot h data practitioners and library developers. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Accelerated DataFrames for all: Bringing GPU acceleration to pandas and Polars - Vyas Ramasubramani URL:https://cfp.scipy.org/scipy2025/talk/EMLLYF/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-Z7AL7K@cfp.scipy.org DTSTART;TZID=PST:20250710T163000 DTEND;TZID=PST:20250710T170000 DESCRIPTION:We illustrate the power and flexibility of a new extension poin t in Xarray's data model: "custom indexes" that allow Xarray users to neat ly handle complex grids\, and enables at least one new data model (vector data cubes). We present a whirlwind tour of specific examples to illustrat e the power of this feature\, and aim to stimulate experimentation during the sprints. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:The brave new world of slicing and dicing Xarray objects. - Deepak Cherian\, Scott Henderson\, Benoît Bovy\, Justus Magin URL:https://cfp.scipy.org/scipy2025/talk/Z7AL7K/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-9QG8PU@cfp.scipy.org DTSTART;TZID=PST:20250710T163000 DTEND;TZID=PST:20250710T170000 DESCRIPTION:Working with data in grids or spreadsheets is great for collabo ration as there are many different tools to view and edit the files. Data science workflows often include packages like openpyxl to create\, load\, edit\, and export spreadsheets that then are shared with others who can us e other tools like Excel\, Google Sheets\, or IDEs to view them. The new P ython in Excel feature as well as the Anaconda Toolbox add-in provides the tools to run Python directly in cells in a spreadsheet\, making it easier for Pythonistas to access and collaborate on code. This talk will introdu ce how these features work\, demo collaborating on Python code in a worksh eet\, and talk about some case studies where these tools have been used to teach and collaborate with Python. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Getting all your snakes in a grid: collaborating and teaching with Python in Excel and the Anaconda Toolbox - Sarah Kaiser URL:https://cfp.scipy.org/scipy2025/talk/9QG8PU/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-EPTPDH@cfp.scipy.org DTSTART;TZID=PST:20250710T163000 DTEND;TZID=PST:20250710T170000 DESCRIPTION:High energy particle (HEP) physics research is going through fu ndamental changes as we move to collect larger amounts of data from the La rge Hadron Collider (LHC). Analysis facilities and distributed computing\, through HTCs\, have come together to create the next pythonic generation of analysis by utilizing htcdaskgateway\, a Dask gateway extension\, allow ing users to spawn workers compatible with both their analysis and heterog eneous clusters in line with authentication requirements. This is enabling physicists to engage with scientific python in ways they had not before b ecause of domain specific C++ tools. An example of htcdaskgateway’s use is Fermilab’s Elastic Analysis Facility. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Enabling Innovative Analysis on Heterogeneous Clusters through HTCd askgateway - Elise Chavez URL:https://cfp.scipy.org/scipy2025/talk/EPTPDH/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-XDFZXG@cfp.scipy.org DTSTART;TZID=PST:20250710T172000 DTEND;TZID=PST:20250710T182000 DESCRIPTION:Lightning talks are 5-minute talks on any topic of interest for the SciPy community. We encourage spontaneous and prepared talks from eve ryone\, but we can’t guarantee spots. Sign ups are at the NumFOCUS booth during the conference. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Lightning Talks - URL:https://cfp.scipy.org/scipy2025/talk/XDFZXG/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-HTHTXS@cfp.scipy.org DTSTART;TZID=PST:20250710T183000 DTEND;TZID=PST:20250710T192500 DESCRIPTION:This BoF aims to host discussion about best practices for maint aining executable tutorials that are reproducible and reliable. The BoF is intended to be a platform to collect tips and tricks of CI/CD practices\, too. \nThe moderators recently put together a repository that builds on t heir experiences of maintaining numerous tutorial repositories https://sci entific-python.github.io/executable-tutorials/ that covers some of the use cases but we are well aware that there are still user scenarios and use c ases that are not well covered.\n\nThe BoF is a complement for both the Te aching&Learning and Maintainers track\, none of the talks in those tracks seem to focus on the technical challenges around tutorials. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Reliable executable tutorials -- CI/CD challenges - Brigitta Sipőc z URL:https://cfp.scipy.org/scipy2025/talk/HTHTXS/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-MZAE3X@cfp.scipy.org DTSTART;TZID=PST:20250710T183000 DTEND;TZID=PST:20250710T192500 DESCRIPTION:Research software engineer (RSE) communities of practice specif ic to a given science are crucial social structures between developers\, m aintainers and users prompting naturally occurring peer mentoring opportun ities\, software improvements through collaborative contributions\, and sh aring of best practices and lessons learned from challenges specific to th at science discipline. Members of such communities benefit from the vast r esources and support available through other RSEs of their own scientific field\, and the users of those software benefit from a more capable and us er-friendly product.\n\nWhile the US-RSE (us-rse.org) advocates for recogn ition of the overall RSE community\, provides individual RSEs with a sense of belonging (e.g.\, inclusivity)\, and provides helpful resources\, it l acks the science specific support possible in more focused communities of practice. This session features short scene-setting presentations\, follow ed by an open panel discussion with leaders of science-specific communitie s of practice for RSEs (e.g.\, Python in Heliophysics Community (PyHC)\, P lanetaryPy\, earthaccess\, and Pangeo) on the benefits of and lessons lear ned from leading those groups in comparison to more general RSE communitie s. Example discussion topics include the benefits of science-specific RSE communities\, development of science-specific software standards\, encoura ging psychological safety\, and community creation and sustainability. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Open-source science-specific Research Software Engineer Communities : benefits and lessons learned - Julie Barnum URL:https://cfp.scipy.org/scipy2025/talk/MZAE3X/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LSJV33@cfp.scipy.org DTSTART;TZID=PST:20250710T183000 DTEND;TZID=PST:20250710T192500 DESCRIPTION:AI\, particularly generative AI\, is rapidly transforming the s cientific landscape\, offering unprecedented opportunities and novel chall enges across all stages of research. This Birds of a Feather session aims to bring together researchers\, developers\, and practitioners to share ex periences\, discuss best practices\, and explore the evolving role of AI i n science. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:AI for Scientific Discovery - Inessa Pawson\, Jonathan Starr\, Cord ero Core\, Anant Mittal\, Carlos García Jurado Suarez URL:https://cfp.scipy.org/scipy2025/talk/LSJV33/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-J9DDE8@cfp.scipy.org DTSTART;TZID=PST:20250711T091500 DTEND;TZID=PST:20250711T100000 DESCRIPTION:After two decades of planning\, Rubin Observatory is finally ob serving the sky. Built to image the entire southern hemisphere every few n ights with a 3.2-gigapixel camera\, Rubin will produce a time-lapse of the Universe\, revealing moving asteroids\, pulsing stars\, supernovae\, and rare transients that you only catch if you're always watching.\n\nIn this talk\, I'll share the “first look” images from Rubin Observatory as w ell as what it took to get here: from scalable algorithms to infrastructur e that moves data from a mountaintop in Chile to scientists around the wor ld in seconds. I'll reflect on what we learned building the data managemen t system in Python over the years\, including stories of choices that impa cted scalability\, interfaces\, and maintainability. Rubin Observatory is here. And it's for you. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Rubin Observatory: What will you discover when you’re always watc hing - Dr. Yusra AlSayyad URL:https://cfp.scipy.org/scipy2025/talk/J9DDE8/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-DTJMTR@cfp.scipy.org DTSTART;TZID=PST:20250711T104500 DTEND;TZID=PST:20250711T111500 DESCRIPTION:As scientific computing increasingly relies on diverse hardware (CPUs\, GPUs\, etc) and data structures\, libraries face pressure to supp ort multiple backends while maintaining a consistent API. This talk presen ts practical considerations for adding dispatching to existing libraries\, enabling seamless integration with external backends. Using NetworkX and scikit-image as case studies\, we demonstrate how they evolved to become a common API with multiple implementations\, handle backend-specific behavi ors\, and ensure robustness through testing and documentation. We also dis cuss technical challenges\, differences in approaches\, community adoption strategies\, and the broader implications for the SciPy ecosystem. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Lessons Learned from Adding Backend Dispatching to NetworkX and sci kit-image - Erik Welch URL:https://cfp.scipy.org/scipy2025/talk/DTJMTR/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-KE7C9X@cfp.scipy.org DTSTART;TZID=PST:20250711T104500 DTEND;TZID=PST:20250711T111500 DESCRIPTION:The SciPy Proceedings () have lo ng served as a cornerstone for publishing research in the scientific pytho n community\; with over 330 peer-reviewed articles being published over th e last 17 years. In 2024\, the SciPy Proceedings underwent a significant t ransformation\, adopting MyST Markdown () and Curvenot e () to enhance accessibility\, interactivity\, and reproducibility — including publishing of Jupyter Notebooks. The new pr oceedings articles are web-first\, providing features such as deep-dive li nks for cross-references and previews of GItHub content\, interactive 3D v isualizations\, and rich-rendering of Jupyter Notebooks. In this talk\, we will (1) present the new authoring & reading capabilities introduced in 2 024\; (2) highlight connections to prominent open-science initiatives and their impact on advancing computational research publishing\; and (3) demo nstrate the underlying technologies and how they enhance integrations with SciPy packages and how to use these tools in your own communication workf lows.\n\nOur presentation will give an overview of the revised authoring p rocess for SciPy Proceedings\; how we improve metadata standards in a simi lar way to code-linting and continuous integration\; and the integration o f live previews of the articles\, including auto-generated PDFs and JATS X ML (a standard used in scientific publishing). The peer-review process for the proceedings currently happens using GitHub’s peer-review commenting in a similar fashion to the Journal of Open Source Software\; we will dem onstrate this process as well as showcase opportunities for working with d istributed review services such as PREreview (). Th e open publishing pipeline has streamlined the submission\, review\, and r evision processes while maintaining high scientific quality and improving the completeness of scholarly metadata. Finally\, we will present how this work connects into other high-profile scientific publishing initiatives t hat have incorporated Jupyter Notebooks and live computational figures as well as interactive displays of large-scale data. These initiatives includ e [*Notebooks Now!*](https://data.agu.org/notebooks-now) by the American G eophysical Union\, which is focusing on ensuring that Jupyter Notebooks ca n be properly integrated into the scholarly record\; and the Microscopy So ciety of America’s work on [interactive publishing](https://elementalmic roscopy.org) and publishing of large-scale microscopy data with interactiv e visualizations. These initiatives and the SciPy Proceedings are enabled by recent improvements in open-source tools including MyST Markdown\, Jupy terLab\, BinderHub\, and Curvenote\, which enable new ways to share execut able research content. These initiatives collectively aim to improve both the reproducibility\, interactivity\, and the accessibility of research by providing improved connections between data\, software and narrative rese arch articles.\n\nBy embracing open science principles and modern technolo gies\, the SciPy Proceedings exemplify how computational research can be m ore transparent\, reproducible\, and accessible. The shift to computationa l publishing\, especially in the context of the scientific python communit y\, opens new opportunities for researchers to publish not only their fina l results but also the computational workflows\, datasets\, and interactiv e visualizations that underpin them. This transformation aligns with broad er efforts in open science infrastructure\, such as integrating persistent identifiers (DOIs\, ORCID\, ROR)\, and adopting FAIR (Findable\, Accessib le\, Interoperable\, Reusable) principles for computational content. Build ing on these foundations\, as well as open tools like MyST Markdown and Cu rvenote\, provides a scalable model for open scientific publishing that br idges the gap between computational research and scholarly communication\, fostering a more collaborative\, iterative\, and continuous approach to s cientific knowledge dissemination. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:SciPy Proceedings: An Exemplar for Publishing Computational Open Sc ience - Rowan Cockett URL:https://cfp.scipy.org/scipy2025/talk/KE7C9X/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-PNSEX8@cfp.scipy.org DTSTART;TZID=PST:20250711T104500 DTEND;TZID=PST:20250711T111500 DESCRIPTION:Neuroscientists record brain activity using probes that capture rapid voltage changes ('spikes') from neurons. Spike sorting\, the proces s of isolating these signals and attributing them to specific neurons\, fa ces significant challenges: incompatible file formats\, diverse algorithms \, and inconsistent quality control. SpikeInterface provides a unified Pyt hon framework that standardizes data handling across technologies and enab les reproducibility. In this talk\, we will discuss: 1) SpikeInterface's m odular components for I/O\, processing\, and sorting\; 2) containerized de pendency management that eliminates complex installation conflicts between diverse spike sorters\; and 3) parallelization tools optimized for the me mory-intensive nature of large-scale electrophysiology recordings. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:SpikeInterface: Streamlining End-to-End Spike Sorting Workflows - H eberto Mayorquin\, Alessio Buccino URL:https://cfp.scipy.org/scipy2025/talk/PNSEX8/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-TQK3T9@cfp.scipy.org DTSTART;TZID=PST:20250711T112500 DTEND;TZID=PST:20250711T115500 DESCRIPTION:The SciPy library provides objects representing well over 100 u nivariate probability distributions. These have served the scientific Pyth on ecosystem for decades\, but they are built upon an infrastructure that has not kept up with the demands of today’s users. To address its shortc omings\, SciPy 1.15 includes a new infrastructure for working with probabi lity distributions. This talk will introduce users to the new infrastructu re and demonstrate its many advantages in terms of usability\, flexibility \, accuracy\, and performance. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:SciPy’s New Infrastructure for Probability Distributions and Rand om Variables - Matt Haberland\, Albert Steppi URL:https://cfp.scipy.org/scipy2025/talk/TQK3T9/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-Z78CFX@cfp.scipy.org DTSTART;TZID=PST:20250711T112500 DTEND;TZID=PST:20250711T115500 DESCRIPTION:Extreme weather events threaten industries and economic stabili ty. NOAA’s National Centers for Environmental Information (NCEI) address es this through the Industry Proving Grounds (IPG)\, which modernizes data delivery by collaborating with sectors like re/insurance and retail to de velop practical\, data-driven solutions. This presentation explores IPG’ s technical innovations\, including implementing Polars for efficient data processing\, AWS for scalability\, and CI/CD pipelines for streamlined de ployment. These tools enhance data accessibility\, reduce latency\, and su pport real-time decision-making. By integrating scientific computing\, clo ud technology\, and DevOps\, NCEI improves climate resilience and provides a model for leveraging open-source tools to address global challenges. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:From Legacy to Leading-Edge: Revamping NCEI Software for the Cloud Era - Sarah Purpura URL:https://cfp.scipy.org/scipy2025/talk/Z78CFX/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-K98LXU@cfp.scipy.org DTSTART;TZID=PST:20250711T112500 DTEND;TZID=PST:20250711T115500 DESCRIPTION:The elasticity of the Cloud is very appealing for processing la rge scientific data. However\, enormous volumes of unstructured research d ata\, totaling petabytes\, remain untapped in data repositories due to the lack of efficient parallel data access. Even-sized partitioning of these data to enable its parallel processing requires a complete re-write to sto rage\, becoming prohibitively expensive for high volumes. In this article we present Dataplug\, an extensible framework that enables fine-grained pa rallel data access to unstructured scientific data in object storage. Data plug employs read-only\, format-aware indexing\, allowing to define dynami cally-sized partitions using various partitioning strategies. This approac h avoids writing the partitioned dataset back to storage\, enabling distri buted workers to fetch data partitions on-the-fly directly from large data blobs\, efficiently leveraging the high bandwidth capability of object st orage. Validations on genomic (FASTQGZip) and geospatial (LiDAR) data form ats demonstrate that Dataplug considerably lowers pre-processing compute c osts (between 65.5% — 71.31% less) without imposing significant overhead s. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Processing Cloud-optimized data in Python (Dataplug) - Universitat Rovira i Virgili (Pedro Garcia Lopez)\, Enrique Molina Giménez URL:https://cfp.scipy.org/scipy2025/talk/K98LXU/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-GXJVBB@cfp.scipy.org DTSTART;TZID=PST:20250711T120000 DTEND;TZID=PST:20250711T125500 DESCRIPTION:Sign up for the CHANCE to give a 5-minute lightning talk by mes saging David Nicholson or Rebecca BurWei on Slack. Or\, show up to the Zoo m on time and we'll take names for the first 5 minutes. Talks will be rand omly selected. Virtual surprises await!\nVirtual and in-person conference attendees welcome!\n\nZoom: https://numfocus-org.zoom.us/j/82704423021?pwd =rJSUmdWwGaqIL8WKY4s6l7B6049rBM.1 \n2025-07-11 12:00 until 2026-07-11 13:0 0 DTSTAMP:20260417T070310Z LOCATION:Virtual SUMMARY:(Exclusively on Zoom) Not Remotely Fun: Virtual Lightning Talks - URL:https://cfp.scipy.org/scipy2025/talk/GXJVBB/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-7MURRS@cfp.scipy.org DTSTART;TZID=PST:20250711T131500 DTEND;TZID=PST:20250711T134500 DESCRIPTION:Open-source projects are intricate ecosystems that consist of h umans contributing in a diverse manner. These contributions are one of the essential elements driving the projects and must be encouraged. The human s behind these contributions play a vital role in constituting the lively and diverse community of the project. Both the humans and their contributi ons must be preserved and handled with utmost care for the success and evo lution of the project.\n\nAs with every community\, certain best practices should be followed to maintain its health\, and certain pitfalls should b e avoided. In this talk\, I’ll share what I have learned from maintainin g the vibrant and wonderful Zarr project and its community over the years. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Learning the art of fostering open-source communities - Sanket Verm a URL:https://cfp.scipy.org/scipy2025/talk/7MURRS/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-SDPLQQ@cfp.scipy.org DTSTART;TZID=PST:20250711T131500 DTEND;TZID=PST:20250711T134500 DESCRIPTION:Real-time machine learning depends on features and data that by definition can’t be pre-computed. Detecting fraud or acute diseases lik e sepsis requires processing events that emerged seconds ago. How do we bu ild an infrastructure platform that executes complex data pipelines (< 10m s) end-to-end and on-demand? All while meeting data teams where they are –in Python–the language of ML!\nLearn how we built a symbolic interpre ter that accelerates ML pipelines by transpiling Python into DAGs of stati c expressions. These expressions are optimized in C++ and eventually run i n production workloads at scale with Velox–an OSS (~4k stars) unified qu ery engine (C++) from Meta. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Real-time ML: Accelerating Python for inference (< 10ms) at scale - Elliot Marx URL:https://cfp.scipy.org/scipy2025/talk/SDPLQQ/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LNWCSE@cfp.scipy.org DTSTART;TZID=PST:20250711T131500 DTEND;TZID=PST:20250711T134500 DESCRIPTION:Would you rather read a “Climate summary” or a “Climate s ummary for _exactly where you live_”? Producing documents that tailor yo ur scientific results to an individual or their situation increases unders tanding\, engagement\, and connection. But\, producing many reports can be onerous. \n\nIf you are looking for a way to automate producing many repo rts\, or you produce reports like this but find yourself in copy-and-paste hell\, come along to learn how Quarto solves this problem with parameteri zed reports - you create a single Python notebook\, but you generate many beautiful customized PDFs.\n\n\n[Slides](https://cwickham.github.io/one-no tebook-many-reports/) DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:From One Notebook to Many Reports: Automating with Quarto - Charlot te Wickham URL:https://cfp.scipy.org/scipy2025/talk/LNWCSE/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-LND7LC@cfp.scipy.org DTSTART;TZID=PST:20250711T135500 DTEND;TZID=PST:20250711T142500 DESCRIPTION:Napari\, an open-source viewer for scientific data\, has an inv iting and well-established community that encourages contribution to its o wn project and the broader bioimage analysis community. This talk will exp lore how napari supports **non-traditional contributors**—especially tho se without formal software development experience—through its welcoming community\, human-centered documentation\, and rich plugin ecosystem. \n As someone with a pure biology background\, I will share my journey into c omputational bioimage analysis and the scientific Python world\, and contr ibuting to napari's community. By sharing my experience writing a plugin and contributing to the core project\, I will show how community-driven pr ojects\, like napari\, lower barriers to entry\, empower scientists\, and cultivate a diverse\, engaged research and developer community. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:From the outside\, in: How the napari community supports users and empowers transition to contribution - Tim Monko URL:https://cfp.scipy.org/scipy2025/talk/LND7LC/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-3KJRWT@cfp.scipy.org DTSTART;TZID=PST:20250711T135500 DTEND;TZID=PST:20250711T142500 DESCRIPTION:The increasing prevalence of AI models necessitates robust mech anisms to ensure their trustworthiness. This talk introduces a standardize d\, PKI-agnostic approach to verifying the origins and integrity of machin e learning models\, as built by the OpenSSF Model Signing project. We exte nd this methodology beyond models to encompass datasets and other associat ed files\, offering a holistic solution for maintaining data provenance an d integrity. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:From Model to Trust: Building upon tamper-proof ML metadata records - Mihai Maruseac URL:https://cfp.scipy.org/scipy2025/talk/3KJRWT/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-CG8P37@cfp.scipy.org DTSTART;TZID=PST:20250711T135500 DTEND;TZID=PST:20250711T142500 DESCRIPTION:Python notebooks are a workhorse of scientific computing. But t raditional notebooks have problems — they suffer from a reproducibility crisis\; they are difficult to use with interactive widgets\; their file f ormat does not play well with Git\; and they aren't reusable like regular Python scripts or modules.\n\nThis talk presents a marimo\, an open-source reactive Python notebook that addresses these concerns by modeling notebo oks as dataflow graphs and storing them as Python files. We discuss design decisions and their tradeoffs\, and show how these decisions make marimo notebooks reproducible in execution and packaging\, Git-friendly\, executa ble as scripts\, and shareable as apps. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:marimo: an open-source reactive Python notebook - Akshay Agrawal URL:https://cfp.scipy.org/scipy2025/talk/CG8P37/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-X7S9PA@cfp.scipy.org DTSTART;TZID=PST:20250711T143500 DTEND;TZID=PST:20250711T150500 DESCRIPTION:PhD students\, postdocs and independent researchers often strug gle when trying to execute code developed locally in the cloud or HPC clus ters for better performance. This is even more difficult if they can't cou nt on IT staff to set up the necessary infrastructure for them on the remo te machine\, which is common in third-world countries. Spyder 6.1 will com e with a whole set of improvements to address that limitation\, from setti ng up a server automatically to easily run code remotely on behalf of user s\, to manage remote Conda environments and the remote file system from th e comfort of a local Spyder installation. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Remote development for students and indie researchers with Spyder - Carlos Cordoba\, C.A.M. Gerlach URL:https://cfp.scipy.org/scipy2025/talk/X7S9PA/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-UEBEUP@cfp.scipy.org DTSTART;TZID=PST:20250711T143500 DTEND;TZID=PST:20250711T150500 DESCRIPTION:[Flyte is a Linux Foundation OSS](https://flyte.org/) orchestra tor built for Data and Machine Learning workflows focused on scalability\, reliability\, and developer productivity. Flyte’s Python SDK\, Flytekit \, empowers developers by shipping their code from their local environment s onto a cluster with one simple CLI command. In this talk\, you will lear n about the design and implementation details that powers Flytekit’s cor e features\, such as “fast registration” and “type transformers”\, and a plugin system that enables Dask\, Ray\, or distributed GPU workflow s. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Dive into Flytekit's Internals: A Python SDK to Quickly Bring your Code Into Production - Thomas J. Fan URL:https://cfp.scipy.org/scipy2025/talk/UEBEUP/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-ZNLF8V@cfp.scipy.org DTSTART;TZID=PST:20250711T143500 DTEND;TZID=PST:20250711T150500 DESCRIPTION:The rapid growth of scientific data repositories demands innova tive solutions for efficient metadata creation. In this talk\, we present our open-source project that leverages large language models to automate t he generation of standard-compliant metadata files from raw scientific dat asets. Our approach harnesses the capabilities of pre-trained open source models\, finetuned with domain-specific data\, and integrated with Langgra ph to orchestrate a modular\, end-to-end pipeline capable of ingesting het erogeneous raw data files and outputting metadata conforming to specific s tandards.\n\nThe methodology involves a multi-stage process where raw data is first parsed and analyzed by the LLM to extract relevant scientific an d contextual information. This information is then structured into metadat a templates that adhere strictly to recognized standards\, thereby reducin g human error and accelerating the data release cycle. We demonstrate the effectiveness of our approach using the USGS ScienceBase repository\, wher e we have successfully generated metadata for a variety of scientific data sets\, including images\, time series\, and text data.\n\nBeyond its immed iate application to the USGS ScienceBase repository\, our open-source fram ework is designed to be extensible\, allowing adaptation to other data rel ease processes across various scientific domains. We will discuss the tech nical challenges encountered\, such as managing diverse data formats and e nsuring metadata quality\, and outline strategies for community-driven enh ancements. This work not only streamlines the metadata creation workflow b ut also sets the stage for broader adoption of generative AI in scientific data management. \n\nAdditional Material:\n- Project supported by USGS an d ORNL\n- Codebase will be available on GitHub after paper publication\n- Fine-tuned LLM models will be available on Hugginface after paper publicat ion DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:Accelerating scientific data releases: Automated metadata generatio n with LLM agents - Tudor Garbulet\, Chirag Shah URL:https://cfp.scipy.org/scipy2025/talk/ZNLF8V/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-XXU9BP@cfp.scipy.org DTSTART;TZID=PST:20250711T153000 DTEND;TZID=PST:20250711T163000 DESCRIPTION:Lightning talks are 5-minute talks on any topic of interest for the SciPy community. We encourage spontaneous and prepared talks from eve ryone\, but we can’t guarantee spots. Sign ups are at the NumFOCUS booth during the conference. DTSTAMP:20260417T070310Z LOCATION:Ballroom SUMMARY:Lightning Talks - URL:https://cfp.scipy.org/scipy2025/talk/XXU9BP/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-AMP9W3@cfp.scipy.org DTSTART;TZID=PST:20250711T164000 DTEND;TZID=PST:20250711T173500 DESCRIPTION:Come share your ideas next year's SciPy. Participants will have an opportunity to sign up to be on next year's organizing committee. DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:SciPy 2026 - Madicken\, Julie Hollek URL:https://cfp.scipy.org/scipy2025/talk/AMP9W3/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-38QDVR@cfp.scipy.org DTSTART;TZID=PST:20250711T164000 DTEND;TZID=PST:20250711T173500 DESCRIPTION:If you have interest in NumPy\, SciPy\, Signal Processing\, Sim ulation\, DataFrames\, Linear Programming (LP)\, Vehicle Routing Problems (VRP)\, or Graph Analysis\, we'd love to hear what performance you're seei ng and how you're measuring. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:GPU Accelerated Python - Katrina Riehl URL:https://cfp.scipy.org/scipy2025/talk/38QDVR/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-FYSAYL@cfp.scipy.org DTSTART;TZID=PST:20250711T164000 DTEND;TZID=PST:20250711T173500 DESCRIPTION:Recent breakthroughs in large language model-based artificial i ntelligence (AI) have captured the public’s interest in AI more broadly. With the growing adoption of these technologies in professional and educa tional settings\, public dialog about their potential impacts on the workf orce has been ubiquitous. It is\, however\, difficult to separate the publ ic dialog about the potential impact of the technology from the experience d impact of the technology in the research software engineer and data scie nce workplace. Likewise\, it is challenging to separate the generalized an xiety about AI from its specific impacts on individuals working in special ized work settings. \n\nAs research software engineers (RSEs) and those in adjacent computational fields engage with AI in the workplace\, the reali ties of the impacts of this technology are becoming clearer. However\, muc h of the dialog has been limited to high-level discussion around general i ntra-institutional impacts\, and lacks the nuance required to provide help ful guidance to RSE practitioners in research settings\, specifically. Sur prisingly\, many RSEs are not involved in career discussions on what the r ise of AI means for their professions. \n\nDuring this BoF\, we will hold a structured\, interactive discussion session with the goal of identifying critical areas of engagement with AI in the workplace including: current use of AI\, AI assistance and automation\, AI skills and workforce develop ment\, AI and open science\, and AI futures. This BoF will represent the f irst of a series of discussions held jointly by the Academic Data Science Alliance and the US Research Software Engineer Association over the coming year\, with support from Schmidt Sciences. The insights gathered from the se sessions will inform the development of guidance resources on these top ic areas for the broader RSE and computational data practitioner communiti es. DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Real-world Impacts of Generative AI in the Research Software Engine er and Data Scientist Workplace - Steve Van Tuyl URL:https://cfp.scipy.org/scipy2025/talk/FYSAYL/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-UZRQJM@cfp.scipy.org DTSTART;TZID=PST:20250711T174500 DTEND;TZID=PST:20250711T184000 DESCRIPTION:In the open-source community\, the security of software package s is a critical concern since it constitutes a significant portion of the global digital infrastructure. This BoF session will focus on the supply c hain security of open-source software in scientific computing. We aim to b ring together maintainers and contributors of scientific Python packages t o discuss current security practices\, identify common vulnerabilities\, a nd explore tools and strategies to enhance the security of the ecosystem. Join us to share your experiences\, challenges\, and ideas on fortifying o ur open-source projects against potential threats and ensuring the integri ty of scientific research. DTSTAMP:20260417T070310Z LOCATION:Room 317 SUMMARY:Towards Robust Security in Scientific Open Source Projects - Juanit a Gomez URL:https://cfp.scipy.org/scipy2025/talk/UZRQJM/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-X8BF9J@cfp.scipy.org DTSTART;TZID=PST:20250711T174500 DTEND;TZID=PST:20250711T184000 DESCRIPTION:Since agent processing take significant time\, what happens to this latency induced if agentic-ai is implemented in existing workflow.\nW hat are the latency challenges ?\nWhat could be key strategies to overcome challenges?\nWhat should we do to change the user expectation.=?\nWhat sh ould be done to maintain/enhance user experience?\nWhat trade-offs should be considers between performance\, latency\, cost etc? DTSTAMP:20260417T070310Z LOCATION:Room 318 SUMMARY:Agentic-Ai and latency implications - Anil Sharma URL:https://cfp.scipy.org/scipy2025/talk/X8BF9J/ END:VEVENT BEGIN:VEVENT UID:pretalx-scipy2025-ZXNVXU@cfp.scipy.org DTSTART;TZID=PST:20250711T174500 DTEND;TZID=PST:20250711T184000 DESCRIPTION:Come join the BoF to do a practice run on contributing to a Git Hub project. We will walk through how to open a Pull Request for a bugfix\ , using the workflow most libraries participating at the weekend sprints u se (hosted by the sprint chairs) DTSTAMP:20260417T070310Z LOCATION:Room 315 SUMMARY:SciPy 2025 Sprint Prep BoF - URL:https://cfp.scipy.org/scipy2025/talk/ZXNVXU/ END:VEVENT END:VCALENDAR