This repository contains a Flask-based API for serving an MLflow model to predict customer churn in a banking dataset. The API loads the latest version of the registered model from MLflow and exposes a REST endpoint for making predictions.

• Fetches the latest version of a registered model from MLflow.
• Exposes a /predict endpoint to make predictions using a trained model.
• Dockerized for easy deployment.
• CI/CD pipeline using Jenkins.

├── app.py                   # Flask application
├── Dockerfile               # Docker container setup
├── requirements.txt         # Python dependencies
├── Jenkinsfile              # CI/CD pipeline setup
└── README.md                # Project documentation

• Python 3.8+
• Docker
• MLflow server with a registered model

pip install -r requirements.txt

python app.py --mlflow_tracking_uri http://mlflow.example.com:8080 --model_name my_model

• Method: POST
• Content-Type: application/json
• Request Body:

{
    "Gender": 1,
    "Balance": 12000.5,
    "NumOfProducts": 2,
    "IsActiveMember": 1,
    "Geography_France": 1,
    "Geography_Germany": 0,
    "Geography_Spain": 0,
    "Age_bin": 3
}

• Response:

{
    "predictions": [0.8]
}

docker build -t bank-churn-api .

docker run -d -p 5000:5000 --name bank-churn-container bank-churn-api

This project includes a Jenkins pipeline (Jenkinsfile) to automate building, tagging, and deploying the Docker container. The pipeline consists of the following stages:

1. Clone Repository
2. Build Docker Image
3. Tag Image
4. Push Image to Registry
5. Run Container

• Flask
• MLflow
• pandas
• scikit-learn
• cloudpickle
• numpy
• psutil
• scipy

This project is licensed under the MIT License.