Whenever you're ready, walk me through how you'd approach this.
How would you monitor a model’s performance after it has been deployed to production?
Explain which service, data-quality, model-quality, and business metrics you would track. Discuss how you would detect drift, handle delayed or missing labels, distinguish model degradation from traffic changes, and trigger investigation, rollback, or retraining.