Automated Fraud Detection & Abuse Prevention Scanner — AI code evaluation for fraud tooling (Dec 2025–Present)
Built and evaluated an automated fraud detection and abuse-prevention scanner that performs behavioral analysis and pattern matching on signals indicating fraud and trust & safety risks. Generated structured JSON fraud risk reports with severity classifications by assessing missing controls, payment anomalies, transaction patterns, CORS-related suspicious behaviors, and data exposure risks. Evaluated AI-generated fraud detection code using Claude Code and validated edge cases, false positives/negatives, and automated system failure modes. • Engineered a Node.js-based fraud detection tool for abuse patterns and trust & safety risks • Produced Critical/High/Medium severity JSON risk reports aligned to industry standards • Assessed AI-generated code in Claude Code to surface edge cases and failure modes • Deployed in production-like environments for fraud prevention testing and trust & safety evaluation