Research Projects Showcase

S3LLM : Large-Scale Scientific Software

I developed S3LLM, a framework designed to simplify the analysis of large-scale software by utilizing the LLaMA 2 model and RAG techniques. This was demonstrated through the analysis of complex programs like the Energy Exascale Earth System Model (E3SM), which has over a million lines of code. I integrated a feature query language (FQL) to enable easy static code analysis without needing advanced programming skills, allowing users to generate FQL queries from natural language. Additionally, I enhanced S3LLM’s metadata interpretation by supporting formats like DOT and SQL and used advanced LLM techniques to improve document analysis for extracting critical information from complex reports.

Get in Touch with Me

Feel free to reach out for collaborations or inquiries.