Tony Mendez

Spearheaded the Indirect Auto Pricing Optimization Solution in Pyomo that leverages desired metrics to create a competitive pricing grid to maximize the profitability and desirability of the indirect auto loans across the multi-billion dollar portfolio. Architect of the UIL Decisioning Optimization Model which significantly cut loss rate while maximizing excess return. Solution provided an intuitive, yet mathematically eloquent way to determine which applicants to accept. Designed an algorithm that enhances the UIL Decisioning Optimization model by developing risk criteria based on key decisioning metrics which efficiently removes riskier cross sections from the optimization process. Lead modeler of Active Rate Light Gradient-Boosting Machine Component Model as part of the Credit Card Valuation Machine Learning Model; prepared data, implemented random forest variable selection, conducted variable imputation, ran correlation reduction, and leveraged lightGBM. Collaborated on the design and implementation of the Consumer Card Line Assignment Optimization Model which transformed the portfolio and increased profitability by tens of millions. Assigns credit lines that maximize profitability, subject to business constraints, utilizing ML projections from the valuation model and incorporates A/B Testing into lines. Directed the end-to-end development of UIL Verification of Income Process which leveraged existing streams of income data to confirm an applicant’s income without having to submit documentation; cutting down on applications subject to underwriting by double digit percents, while simultaneously reducing risk in the portfolio.