How would you design a monitoring system for detecting regressions after model deployment for Autonomous Solutions?
