AI-Oriented Behavioral (AI BQ, NEW)

New behavioral round in the post-April-16 AI-Native loop. Standard STAR-shaped questions reweighted toward AI-tool collaboration, prompt judgment, and recognizing AI failure modes. Requirements A 45-min interviewer-led BQ that adds AI-specific prompts on top of the standard conflict / failure / leadership lineup. Reported question themes: A time you collaborated with AI tools (Claude, GPT, Cursor, etc.) on a real project — how did you orchestrate it? A time you caught an AI tool producing wrong / misleading output — how did you recognize and correct it? A time you decided NOT to use AI for a task — what tipped the decision? A time prompt-tuning was a bottleneck — what did you change and what did you learn? Confirmed Meta-specific prompt themes (from May 2026 candidates): AI-Driven Impact (a story where AI tooling materially changed the project outcome) and Continuous AI Learning (how you keep up with the moving frontier). Standard backbone still present: most-impactful project, conflict resolution, feedback you found hard to hear, failure + recovery, mentoring. Notes The AI-oriented prompts coexist with classic BQ; don't drop the traditional stories. Confirm role-specific framing — MLE/RS candidates report deeper LLM-eval / agent-deployment questions than SWE candidates. This round, like other April-16 additions, is still mutating. Some candidates report no AI-specific BQ questions despite being scheduled into the new loop. The "how do you use AI in your day-to-day work" prompt now also appears in phone screens, not only the onsite BQ — have a crisp, specific answer (which tools, for which tasks) ready early in the loop. Preparation Prep at least 2 dedicated STAR stories about AI collaboration: (1) end-to-end orchestration with a concrete decision, (2) catching an AI mistake and the verification process that surfaced it. Have a no-AI counterexample ready (the right call was to do it yourself or with a teammate, not Claude). Re-rehearse the classic six (impact, conflict, feedback, failure, mentor, push-back); they remain weighted. Use a private prompt-and-response log over the next 2-4 weeks to mine fresh stories.