Vehicle Policy Lapse Prediction
Built an end-to-end machine learning system that identifies customers at high risk of policy lapse to support proactive retention strategies before policy renewal.
- Developed a recall-focused classification workflow using feature engineering and threshold optimisation
- Evaluated multiple models and selected an interpretable solution aligned with business objectives
- Integrated MLflow, FastAPI, and Docker to explore production-oriented ML workflows
Python · Scikit-learn · MLflow · FastAPI · Docker
