Your question is Monitor 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).
Explain your strategy for monitoring model drift in a production environment.
Cover the signals you would monitor, how you would detect meaningful drift, how delayed or missing labels affect monitoring, and how alerts should trigger investigation or rollback. Discuss data quality, training-serving skew, segment-level analysis, and how you would distinguish real drift from normal seasonality.