4:00 pm ET
“The Facility Is the Context: Agentic Computing Infrastructure for Analysis at Scale”
with Dr. Giordon StarkLarge language model agents are becoming useful for software development and scientific analysis, but making them useful on shared computing infrastructure requires more than the model itself. Agents need reliable access to the software, data, workflows, and identities that researchers already use.
This talk introduces the Model Context Protocol (MCP) as a practical interface between AI agents and scientific software. I will briefly discuss how MCP can expose existing Python libraries, databases, experiment services, Jupyter environments, and batch systems, and then show how these pieces come together in an MCP platform developed for the UChicago Analysis Facility.
The platform provides a common gateway for MCP services while handling authentication, authorization, and credential brokering separately from the individual services. This allows agents to work with resources such as Rucio, Jupyter, filesystems, and HTCondor without directly handling users' long-lived credentials.
Using examples from the ATLAS Analysis Facility and work extending this approach across experiments, I will discuss what works today, the infrastructure challenges that remain, and what an MCP-based approach could mean for making agentic computing portable across scientific facilities.
Meet Dr. Giordon Stark
Dr. Giordon Stark is a Deaf project scientist, experimental particle physicist involved with the ATLAS collaboration at the Santa Cruz Institute for Particle Physics, at UC Santa Cruz. He earned his PhD in Physics from University of Chicago in 2018, and a B.S in Physics from Caltech in 2012. Giordon's research focuses on looking for signs of physics Beyond the Standard Model with a particular interest in Electroweak Supersymmetry and hadronic final states. He is also passionate about boosted object reconstruction, jet substructure, pile-up mitigation techniques, designing robust hardware triggers, the intersection of particle physics & machine learning, and more! Through these efforts, Giordon has also made core contributions to improving the communication of physics results between the ATLAS Collaboration and particle physics theorists and phenomenologists. He also led a group of physicists within the experiment to combine and summarize the results of many BSM searches, providing key insights for the future of the successful ATLAS Beyond the Standard Model search program. These results will be made public later this year. When Giordon is not busy trying to prove the existence of new physics, he can be found in the kitchen proving sourdoughs, baking pavlovas, and anything else he can get his hands on.
Recent Events
#52 Dealing with AI Security ‘findings’ — a practical guide to not throwing your laptop in a lake
Holden Karau — October 28, 2026
For anyone maintaining an open source project, it's likely only a matter of time until you start to receive largely AI driven security reports in addition to the AI driven pull requests. Even those who don't maintain open source projects, seeing what's happening behind the scenes in OSS projects will help you understand why your favourite OSS project might seem to have a huge number of "[MINOR][FIX]" commits instead of cool new features right this second.
Historically, we've handled security reports with more caution than pull requests since rejecting a correct security report can have large negative impacts, and the effort required to find a security problem in a project was high.
Now, for better or worse, the effort being high part has changed; so how do we change our responses? This talk will use Holden Karau's personal experience with the Spark 3.5.9/4.0.4/and 4.1.3 releases to look at how to handle a deluge of (non-opt-in) security reports from external developers; as well as how to handle opt-in security reports from select AI labs (if you are granted such a "privilege.") We'll also look at why these AI security reports leave a lot to be desired and hopefully give you some steps to avoid just giving up and throwing your laptop in the nearest large body of water.
Watch on Youtube#51 Join the Forge: Contributing Your First Package to conda-forge
Travis Hathaway — September 30, 2026
conda-forge is a thriving packaging ecosystem with well over 30,000 packages and hundreds of active package maintainers. It's these package maintainers that are the lifeblood of this ecosystem. In this talk, I cover how you can contribute your own packages, why this is beneficial and even more ways to get involved with our community. During the talk, I also briefly cover the history of conda-forge and peak behind the currents to show how its maintainers ensure stability for its users.
Watch on Youtube#50 Nebi: Environment management for teams
Dharhas Pothina — August 26, 2026
Reproducibility of your software environments is often treated like an afterthought. By the time you’re thinking about it, it’s already too late.
We encountered this time again with clients and built a solution in the form of a conda-store which introduced versioning, role-based-accessed control, and most importantly: reproducibility, to conda environments on platforms like JupyterHub.
While effective, the project required heavy infrastructure, had slow environment solves, and was tied to a server-first world of global conda environments. Meanwhile, tools like Pixi and uv have transformed Python packaging with lockfiles and project-oriented workflows, improving reproducibility, but only for individual developers.
In this presentation, we introduce Nebi, a new open-source tool that brings multi-user environment management to the modern Python packaging era, supporting both project-oriented and shared global environments with versioning, role-based access control, and sharing via publishing through OCI registries as well as standalone servers.
Watch on YoutubeAbout Us
At Don’t Use This Code, we want to create a unique opportunity to see Python succeed and thrive within the National Labs! We propose creating a new resource for scientists, researchers, and technical staff to support their use of Python and to build a strong, lasting community for Python users within the Department of Energy National Labs. Disclaimer: The Python Exchange is an independent group of Python enthusiasts who wish to see the use of Python and open-source computing thrive within the National Lab system. This group is not sponsored by or affiliated with the Department of Energy.