Customer Churn MLOps

An end-to-end customer churn prediction project built to demonstrate practical ML and MLOps skills.

Stack

Python, scikit-learn, MLflow, FastAPI, Docker, pytest and GitHub Actions (next step).

Setup

python -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install -r requirements.txt

Download the IBM Telco Customer Churn CSV and place it at:

data/raw/WA_Fn-UseC_-Telco-Customer-Churn.csv

Train

python -m src.train
mlflow ui

Open MLflow at http://localhost:5000 to compare runs and metrics.

Serve the model

uvicorn api.main:app --reload

Open interactive API documentation at http://localhost:8000/docs.

Test

pytest