Major project
Logistics & Supply Chain
Delivery-time prediction service in production
Build a delivery-time model and deploy it as an API a tracking page can call, with tests and monitoring.
The brief
A tracking page should show a reliable delivery time that updates as conditions change. Build the delivery-time model, then deploy it as an API with tests, versioned models and monitoring of prediction error as real delivery times arrive. Good looks like: a deployed API, a release gate on accuracy, a canary release, and an error dashboard.
What you will build
- Build and evaluate the delivery-time model
- Register models in MLflow
- Serve predictions from FastAPI in Docker
- Add tests and a release gate
- Release a new version as a canary
- Monitor prediction error against actual times
Free, open data to build it with
- Amazon delivery dataset ↗ Kaggle
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.