Your question is Evaluating Retrieved Context. 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).
Explain your process for evaluating the relevance and quality of retrieved context in a production pipeline. Cover offline evaluation with labeled examples, online monitoring, retrieval and generation metrics, data-quality checks, sampling strategy, drift detection, and how evaluation results trigger pipeline changes. Discuss how you would preserve experiment lineage and distinguish retrieval failures from downstream answer-generation failures.