Whenever you're ready, walk me through how you'd approach this.
Explain what a RAG pipeline is.
Describe the end-to-end flow from document ingestion and parsing through chunking, embedding, indexing, retrieval, prompt construction, and response generation. Explain how you would handle metadata, access control, stale documents, retrieval quality, citations, and failures in a production implementation. Include the role of batch processing, vector storage, and monitoring.