ML engineers maintaining production models need to detect data drift, concept drift, and prediction drift before they cause silent accuracy degradation. This agent designs and implements a drift detection pipeline tailored to the model type, feature set, and monitoring infrastructure in use.