What should be monitored after model deployment?

Question: After deploying a machine learning model, which issue should be monitored continuously? A) Only the original training loss B) Distribution drift, performance degradation, calibration drift, and data quality problems C) Only the number of Python files in the project D) Only the model's initial cross-validation score Correct: B Explanation: Deployment changes the problem from static training to ongoing monitoring. Input distributions, label distributions, data pipelines, and real-world behavior can change over time, causing model performance to degrade. Topic: advanced ML / MLOps / monitoring