The Problem
An APAC-focused neobank needed to refine its underwriting process for personal loans to minimize non-performing loans (NPLs) while maintaining a competitive approval rate.
Our Solution
Our team engineered and deployed a suite of credit risk models, including Probability of Default (PD) and Loss Given Default (LGD) predictors. We integrated SHAP for model explainability to meet regulatory requirements and built a continuous monitoring dashboard in MLflow to track model drift and performance.