An end-to-end MLOps pipeline that predicts NYC taxi trip duration on 1.45M trips, built as a 3-person group project for the PG Certificate in AI & ML at BITS Pilani.
Covers versioned data and features (DVC), model comparison and experiment tracking (MLflow), a containerised FastAPI prediction service (Docker), automated tests (GitHub Actions), and drift monitoring with an automatic retraining decision. Outlier filtering cut typical-trip error by 32% (MAE 281 s → 190 s, R² 0.78).
0 likes · 0 views · Nov 12, 2026
End-to-end MLOps pipeline predicting NYC taxi trip duration: DVC, MLflow, FastAPI, Docker, CI and drift monitoring. BITS Pilani group project.
