Multi-agent LangGraph RAG for financial Q&A — 72.7% on FinanceBench under κ=0.932 calibrated judge. RBAC at the vector layer, multi-party HITL on high-stakes answers, self-hosted LLM observability. pip install financebench-rag-agent
A ground-up, provider-agnostic toolkit for building, benchmarking, and rigorously evaluating Retrieval-Augmented Generation systems — structure-aware PDF ingestion, five retrieval strategies, and RAGAS end-to-end evaluation, wrapped in a pip install-able CLI and a Streamlit dashboard.
