Senior Lead Quantitative Analytics Specialist
Current♦ Supervised team members in applying machine learning techniques applying random forest, GBM, and explainable machine learning techniques such as EBM, and GAMI_NET to one of the prime card transition models in python. The different machine learning techniques were benchmarked the existing model in terms of the feature selection including interactions and model fitting. ♦ Led in the Prime Card transition model development. The model was a complex model system, which comprised several components including the balance and unit transition among different delinquency state, CECL EAD, and recovery.♦ Led in developing the Vintage Roll Rate model for the Card Assets and Retail Services portfolios. I was in charge of the vintage roll rate adjustment to mitigate the negative model impact from the Covid-19 pandemic. The adjustment tool was greatly appreciated by the business owners and upper management. ♦ Led in loan level survival default model development for Card Assets portfolios including Prime Card, Retail Services, Financial Card, and Dillard’s.♦ Supervised my team members in developing a discounted cash flow model for Card Assets Troubled Debt Restructuring (TDR) portfolio. My team delivered the models on time and provided additional analytics in supporting CMoR's validations. The model received zero severity 1 and 2 risk findings from CMOR. ♦ Developed a short-term roll rate model covering all the unsecured portfolios with unified modeling methodology and approaches. I also addressed numerous outstanding risk findings and concerns surrounding the existing models.♦ Contributed in building the Prime Card Pre-provision Net Revenue (PPNR) models and collaborated with the Home Mortgage modeling and Finance team. ♦ Provided thorough analytics on CCAR/MCST/CECL/BLF model output and presented the results to management/partners.
