• Designed and formulated transaction monitoring (TM) rule typologies and detection criteria on SQL for consumers, commercial businesses, and merchants; established thresholds to flag fraudulent transactions
• Boosted fraud-detection accuracy by 14% by optimizing TM rules through statistical analysis techniques
• Managed team of five to develop origination model for bank’s credit cards portfolio running generalized additive models in python, improving acquisition rate by 15% and profit by $5M
• Leveraged machine learning (GBDT) to construct revenue and attrition models; supported client in devising profitable strategies and bolstered revenue by 10% ; received extraordinary impact award
• Built scoring models factoring in bureau attributes, past behavior and performance data to predict credit card account behavior minimizing likelihood of defaults by 12%
• Coached 18 new analysts through three in-house training workshops on FICO’s modeling techniques
• Delivered 10%-25% reduction in credit losses by employing logistic regression and segmented ensemble models
(SEM) to build debt collection scorecards
• Collaborated across multi-regional teams to upgrade legacy Python scripts for scorecards; increased
computational efficiency of existing scripts by 16%
• Tailored FICO’s proprietary risk models for 10+ small businesses and enhanced Gini coefficient by at least 5%
No Licenses or Certifications Found.