39Our feature's behaviour changed last week although we shipped nothing. How can that happen, and what is your policy on model versions?▼hardNewAnthropicOpenAI◆ premiumAliases move, providers change infrastructure and retired models get re-routed. Pin versions in production, watch continuously and treat every model change as a release.Practise this question →Read answer →
14Support tickets say the AI got worse this month, but every eval still passes. What do you do?▼hardNewNotionPerplexityIntercom◆ premiumGreen dashboards and angry users at the same time. The strong answer treats the complaints as evidence about the eval, not noise to explain away.Practise this question →Read answer →
14Most users never complain when the AI is wrong. What metrics would tell you it is failing anyway?▼mediumNewKlarnaRevolutStripe◆ premiumThe most expensive AI failures are the confident wrong answers that users accept. This question tests whether you can find their footprints in behaviour data.Practise this question →Read answer →
22Walk me through the lifecycle of a production model, from data to retirement. What does a PM own at each stage?▼easyNewGoogleMetaUber◆ premiumA fundamentals question for platform and model PM roles. The interviewer wants the stages in order and, at each, the decision a PM is responsible for.Practise this question →Read answer →