Machine Learning Zoomcamp 2025 – Modern deployment workshop using FastAPI, uv, and Fly.io.

🎯 Goal of the Module

Transition from a trained model in a notebook to a fully deployed web service.

This module focused on the practical steps of serving, containerizing, and deploying a model.

What I learned:

• Preparing data for ML model
• Saving and loading models with pickle
• Serving predictions via FastAPI
• Validating input with Pydantic
• Managing environments with uv
• Containerizing with Docker
• Deploying to the cloud using Fly.io

🧠 Project Context/ Data

The project predicts customer churn — whether a telecom customer will cancel their subscription.

The goal was to expose this trained model as a REST API that can be queried in real time.

The link for the dataset of this project is at :

https://www.kaggle.com/code/farazrahman/telco-customer-churn-logisticregression