A production-ready FastAPI application that predicts whether a loan applicant is a Good or Bad credit risk based on structured financial and demographic data. This project demonstrates how to deploy a trained ML model behind a RESTful API interface.

The API takes in customer input via JSON and returns a prediction based on a Decision Tree Classifier trained on credit data. The model is serialized using pickle (tree_Accuracy_based.pkl) and loaded into the app during runtime.

Built with FastAPI — a blazing-fast, modern Python web framework — this API is ideal for deploying ML models in real-world applications.

• ✅ Input validation using Pydantic models (bank.py)
• ✅ Model loaded dynamically via Python’s pickle
• ✅ Clean JSON endpoints for easy integration
• ✅ Automatic Swagger Docs at /docs
• ✅ Human-readable prediction response

1. The user sends a POST request with financial data in JSON format.
2. FastAPI + Pydantic validates and parses the input.
3. The decision tree model makes a prediction (1 = Good, 0 = Bad).
4. The API returns a friendly response like:

   {"Prediction": "Good"}