Practice Project: Production RAG System

Build a RAG system that evaluates itself

🔑 Key Concepts

Project goal — End-to-end RAG system with document ingestion, hybrid search, reranking, citation, and automated evaluation.
Requirements — Ingest 50+ documents, hybrid search, reranker, citation output, RAGAS evaluation, streaming API.
Extensions — Multi-modal RAG (images + text), conversational memory, query rewriting, A/B testing of retrieval strategies.
Success criteria — Faithfulness > 0.85, context recall > 0.80, sub-2s latency, citations verified.

💡 Practice: Try implementing each concept yourself before moving on. Reading about RAG and building RAG are very different things.