npogeant/flight-delay

Implements automated retraining and model drift detection through orchestrated workflows—Prefect schedules training and monitoring flows that continuously validate model performance against reference datasets using Evidently, while BentoML containerizes the best-performing model as a REST API that Streamlit consumes for predictions. Uses S3 for centralized data storage and MLflow's cloud-hosted tracking server to maintain a production model registry that automatically promotes superior models based on precision metrics.

No commits in the last 6 months.