Your question is Design High-Update RAG Pipeline. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you design a RAG pipeline to handle high-frequency document updates while maintaining low latency?
Explain how documents are ingested, incrementally indexed, retrieved, and used for grounded generation. Address consistency, stale results, caching, failure handling, hallucination control, prompt injection, and how you would evaluate both freshness and answer quality.