Real-time Risk & Compliance Infrastructure
Architected core risk and compliance infrastructure for a fintech platform handling $500M+ in annual transaction volume.
Context & Problem
A rapidly scaling fintech needed to move from manual risk checks to automated, real-time decision systems while staying compliant with regulatory requirements.
Role & Responsibilities
VP Engineering. Owned the risk platform roadmap, built and scaled the team to 8 engineers, drove architecture decisions.
Architecture & Key Decisions
Built a multi-stage decision pipeline: feature extraction → rules engine → ML models → final decision. Used Kafka for event streaming, stored decisions in time-series DB for auditability.
Impact & Outcomes
Reduced fraud by 85%, enabled 10x growth in transaction volume, achieved SOC2 compliance, cut manual review overhead from 20% to 2%.
What I'd Do Differently
Auditability and reversibility are as important as performance. Design for the day you need to explain a decision to regulators. Hybrid rules + ML is often better than pure ML.