What candidates say the round looked like, summarised in our words, each linked to where it came from. Intuit hasn’t confirmed these details unless the source says “Official page”.
75-minute staff technical interview with a coding part and an AI Q&A part. The AI questions came mostly from the candidate’s own recent AI project: database choices, resilience patterns such as rate limiting and circuit breakers, the difference between an LLM and an agent, guardrails against hallucination, why use an MCP server instead of giving the LLM APIs directly, and user versus system prompts.
Staff technical screen split into about 35 minutes of LeetCode-style coding and 35 minutes of AI questions. The AI half was spoken questions, and this candidate found them basic: what RAG is, and when you should and shouldn’t use AI.