Your question is Manage Production Model Drift. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
You've shipped a model that was performing acceptably at launch, but over time its production behavior starts to change. The team wants a clear plan for detecting drift, understanding whether it is feature shift, label shift, or score drift, and deciding when to recalibrate, retune thresholds, retrain, or roll back.
How would you manage model drift in a production AI system?