Your question is Designing a 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 retrieval-augmented generation (RAG) pipeline for an AI application at Delta Air Lines?
Explain the practical architecture from document ingestion through retrieval, reranking, generation, citations, and response delivery. Discuss how you would evaluate answer quality and retrieval quality offline and online, while addressing hallucination, prompt injection, access control, latency, and cost. Include the assumptions, design choices, and implementation details you would validate before production.