data-drifts

This project builds a production-grade ML pipeline to classify Near-Earth Objects (NEOs) as hazardous or non-hazardous. It automates data ingestion, preprocessing, model training, monitoring, and drift detection using GitHub Actions, PostgreSQL, MLflow, DAGsHub, and Grafana.

A system for monitoring statistical data distribution shifts in distributed settings

pip package for data analysis stability evaluation against small data change.

Projet d'apprentissage des pratiques MLOps : pipeline complet de prédiction de prix immobiliers, du suivi d'expériences (MLflow) au serving (FastAPI), avec monitoring de dérive (Evidently) et CI/CD