The Department of Energy has selected a University of Alabama-led team for Phase 1 funding under its Genesis Mission, a national initiative that pairs AI with supercomputing and advanced scientific instruments to accelerate discovery. The project, one of 278 chosen from more than 5,000 proposals, will build an AI tool to automate data analysis for the Compact Muon Solenoid (CMS) experiment at CERN's Large Hadron Collider.
The CMS detector collects about one petabyte of data per second - enough to keep a team of scientists busy for years. That backlog slows the pace of research in high-energy physics, where researchers sift through massive datasets to identify rare particle interactions. The UA-led effort aims to compress that timeline by automating the software and workflow tasks that support analysis.
Automating the routine work of physics
Dr. Konstantin Matchev, Endowed Shelby Distinguished Professor of physics and astronomy at The University of Alabama, is the project's lead principal investigator. He frames the work in practical terms.
"This project is about giving scientists more time to do science," Matchev said. "By automating the routine - but highly complex - software and workflow tasks that support high-energy physics analysis, we can help researchers move more quickly from massive datasets to meaningful discoveries."
The project builds on existing collaborations between UA and the Fermi National Accelerator Laboratory in Chicago. Dr. Sergei Gleyzer, a professor of physics and astronomy and chief science officer of UA's High Performance Computing and Data Center, is one of the UA scientists working with the CMS experiment. In 2025, the team developed a pilot framework that applied large language models to research tasks in particle theory; that prototype will serve as the backbone of the new project.
Scaling beyond particle physics
UA's CMS group has been among the leaders in applying machine learning to high-energy physics. The Phase 1 Genesis project will expand the existing prototype into an intelligent, autonomous tool for processing the CMS data streams, with the goal of making the approach scalable to other data-heavy disciplines.
UA leads the joint project with Fermilab. Matchev and Gleyzer lead for UA, and Dr. Stephen Mrenna serves as project lead at Fermilab. The University of Alabama High-Performance Computing and Data Center will supply computing and AI resources, with additional support from Google and NVIDIA.
"The University of Alabama has one of the largest university-based Compact Muon Solenoid groups in the Southeast, and this award reflects the kind of national and international research leadership our faculty bring to some of science's most ambitious questions to create new human knowledge that is critical to better lives," said Dr. Bryan Boudouris, UA's vice president for research.
For researchers in data-intensive fields, the project offers a concrete model for how AI for Science & Research can shorten the path from raw data to published results. The same pattern - large language models trained to handle domain-specific workflow tasks - could apply to genomics, climate modeling, or materials science, where similar bottlenecks exist.
The broader lesson for Research teams is that the bottleneck in modern science is often not data collection but data processing. Automating the routine analytical steps, as this project proposes, lets investigators spend their time on interpretation and hypothesis generation instead of pipeline management.
Your membership also unlocks:
