FERMI-LLM: Language Models for Gamma-Ray Astrophysics
From a plain-English request to a verified Fermi-LAT analysis.
Analyzing data from NASA's Fermi Large Area Telescope requires expert knowledge of specialized tools. Funded by NASA's Fermi Guest Investigator Program, FERMI-LLM builds a language-model platform that turns a plain-English request into a real Fermi-LAT analysis: it writes the FermiPy configuration and Python driver, checks them against guardrails, runs exactly what the scientist approves, and shows every step of the result.
The platform is designed to be extended by the community: models, rules, validators, data sources and export formats are modules that can be added or replaced. Undergraduate researcher Weber Whelan built an LLM service for Fermi-LAT analysis as a departmental honors thesis (2026), and a Clemson-funded postdoctoral position in AI-enhanced computational astrophysics complements the project.