Your question is Detect Data Drift in Production. 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).
How do you test for data drift in a production machine learning model, and what statistical tests do you rely on?
Explain a practical monitoring approach for numerical and categorical features. State the null hypotheses, show how you would select tests, define alert thresholds, and distinguish statistically significant drift from operationally meaningful drift. Address multiple features, repeated monitoring, missing values, and the actions taken after an alert.