At HDFC Bank Limited, I led the development and deployment of a behavior-segmented mule detection model across savings accounts, improving incremental fraud capture by 57% for a previously unidentified segment.
I also delivered a LightGBM-based reactive mule identification model for current accounts, from data integration and feature engineering through batch scoring deployment. It increased detection efficiency 200x and improved recall by about 2% over the legacy model.
Before HDFC Bank, I worked in Digital and AI at TVS Motor Company, where I developed a product recommendation engine for its customer-facing vehicle app. The recommendations improved average accessories cart value by about ₹250.
At TVS Motor Company, I also worked on demand forecasting, survey analysis, and Power BI dashboards. My work included automating reporting processes and developing insights that supported customer experience and dealership quality improvements.

