How do you build a ground-truth dataset to evaluate if your RAG system is actually improving over time?

Explain the "Self-RAG" or "Corrective RAG" pattern. How does the model decide if it needs to go back and retrieve more data?

Why might you use a "Cross-Encoder" re-ranker after your initial vector retrieval? What is the trade-off in terms of latency?

If a user asks a vague question like "Tell me more about that," how do you ensure the retriever finds relevant documents from past conversation history?