AI Engineer | Building ML pipelines, FastAPI services, and real-time analytics systems
Pinned Loading
-
rag-degradation-paper
rag-degradation-paper PublicEmpirical study of multi-turn retrieval degradation in agentic RAG systems. 150 sessions, 550 turns from a production math tutoring deployment. Pearson r = -0.283, p < 0.0001.
TeX
-
Math-Professor-Ai
Math-Professor-Ai PublicProduction agentic RAG system — decomposes math problems into sub-queries, reasons step-by-step with growing context window, and logs retrieval degradation across turns using Google Gemini 2.5
TypeScript
-
-
-
-
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.