Data Scientist - Fraud Risk
Menlo Park, Ca, Us
In this role, I specialized in ensuring the safety and integrity of Meta's Ads and Commerce platforms by building data-driven solutions for fraud prevention. Key accomplishments include:* Automated decision-making: Built a predictive model, automating 10% of daily appeal tickets with 95% precision, reducing manual intervention.* Revenue optimization: Unblocked $60M in incremental revenue by A/B testing and recommending a novel ticket prioritization model, transitioning from FIFO to a more efficient review process.* Anomaly detection: Reduced fraud detection time from days to hours by implementing an anomaly detection system that flagged trends in real time, improving response times and reducing operational risk.* Fraud detection models: Developed models to identify and enforce on account takeovers, fake accounts, and fraudulent transactions across various user journey stages.* Infrastructure development: Designed and built a scalable system for appeal evaluation, defining metrics, logging features, and creating pipelines to monitor system health and improve decision accuracy.