RAGLens

A live RAG pipeline with LLM-as-judge evaluation. Ask a question — the system retrieves relevant chunks from a knowledge base, generates a grounded answer, then scores both retrieval quality and answer quality automatically using a second AI call.