rag-eval-pipeline
SourceRetrieval-augmented QA over a PDF corpus, built around a from-scratch recall@k / hit-rate@k / MRR evaluation harness — exposed as MCP tools and orchestrated as a LangGraph agent (Qdrant, FastAPI, AWS S3 + Bedrock).
Retrieval-augmented QA over a PDF corpus, built around a from-scratch recall@k / hit-rate@k / MRR evaluation harness — exposed as MCP tools and orchestrated as a LangGraph agent (Qdrant, FastAPI, AWS S3 + Bedrock).