San Jose, California, United States
Modernizing PayPal's USA & UK $2B Business Loans platform—identity, underwriting, and funding decisioning systems.
Platform Modernization & ML Model Deployment
• Replaced a 7-year legacy Salesforce/Swift stack with a unified decisioning platform supporting identity, underwriting, and funding flows.
• Owned end-to-end rollout of Credit Activation probability, Acquisition fraud and Sales-based Revenue prediction models: documented 3000+ Bureau features, internal credit record features, scaled analytical variables, and implemented audit-grade quality testing.
• Built a shadow-mode audit system that simulates bureau pulls for applicants to validate new models without a full E2E user flow.
• Introduced identity-linking logic using tax ID and SSN to connect non-PayPal applicants with existing PayPal loan histories, preserving risk continuity without restricting new customers.
Infrastructure & GTM
• Stood up 6 new bureaus by designing standardized API mapper—reduced integration time from weeks to days, enabling rapid GTM.
• Negotiated dynamic API contracts with TransUnion, activating only required data packages per API request and cutting costs by 25%.
• Diagnosed excessive OFAC scan noise (30 FPs monthly) and deployed an in-house matching API, eliminating recurring compliance bugs.
Real-time Graph Data Systems
• Enriched PayPal's real-time identity graph by combining offline credit data with live account events, enabling real-time exposure scoring for linked PayPal accounts with their loan accounts' history.
• Designed data schema connecting loans and accounts through shared identifiers such as email, phone, and beneficial owner—establishing unified fraud lineage across PayPal credit products.
