ML/AI
Credit Score Explainability & Simulation Platform
Production-grade web platform for real-time credit score simulation and explainability, enabling transparent risk decisions and trust in ML-driven assessments.
I led the end-to-end development of a production-ready web platform that makes company credit scores transparent and interactive for business users. The application exposes model behavior through feature importance, scenario-based simulations, and intuitive visualizations so users can test ‘what-if’ changes and understand why a score looks the way it does. By turning opaque model outputs into explainable, auditable insights, the platform increased trust in credit risk models and supported broader adoption across risk, sales, and product teams.

Architecture and Stack
PythonFastAPIStreamlitscikit-learnXGBoostPostgreSQL