Skip to content
ML Academy

RAG and agents

Interview question

  • seniorlevel
  • 35 minsuggested time
  • 2Applied AI, ML engineer

Whiteboard case. Design an internal RAG assistant for a company of 20,000 employees. Documents live in many systems (wiki, tickets, drive, HR) with their own permissions; some are confidential and permissions change daily. The assistant must never reveal content a user could not open in the source system. Walk through ingestion, indexing, retrieval, generation and evaluation.

Tests: Security-aware RAG design: access control at retrieval time, permission freshness, prompt-injection exposure from retrieved text, and how to evaluate leakage.

day 48day 53day 52day 55

Sign in to practise: answers are saved before model answers are shown.