RAG Pipeline (Design + Evaluation Metrics)

Problem Overview

Design a Retrieval-Augmented Generation (RAG) system end to end: ingestion, chunking, embedding, vector retrieval, and augmented generation — plus how you'd evaluate it. Reported in two forms: (1) an onsite system-design interview requiring the full pipeline design AND a discussion of evaluation metrics; (2) a 4-hour take-home OA where RAG…

• 3 candidate-reported follow-ups — the exact probes interviewers asked, with the trigger for each
• The rest of the problem statement — full requirements, constraints, and edge cases
• Approach and trade-offs — what passing candidates did, and the mistakes that sink people