ML Model Evaluation & Monitoring Framework

A Python-based framework for continuous evaluation and monitoring of ML models in production. The system runs scheduled evaluation jobs that measure model accuracy, inference latency, and data drift using statistical tests (KS, PSI, Jensen-Shannon divergence). Results are stored in a time-series format and visualized through configurable dashboards. The framework supports A/B comparison between model versions, generates automated alerts when drift thresholds are exceeded, and integrates with existing ML pipelines via a lightweight SDK. Used internally to monitor multiple production models with sub-minute evaluation cycles.