Dmitriy Petrakov

Designed and built anti-fraud architecture from scratch using IP intelligence, device fingerprinting and behavioral analysis based on MaxMind. Calibrated risk-scoring thresholds, reducing false positives from 15% to 2-4%. Detected Playwright, Puppeteer and Selenium abuse via DOM-click and JS-injection anomalies, cutting mass bot registrations by 32%. Built account and payment protection with anti-detect browser and proxy detection; reduced manual review by 34% and disputed $80,000+ chargebacks.

Implemented AI-generated and deepfake document checks across KYC/KYB, reducing automated verification costs by 25%. Ran AML/KYT monitoring, EDD reviews, blockchain analysis, OSINT and SoF/SoW checks; identified 20-25 connected clusters with 500+ users attempting mixer cash-outs and escalated cases to law enforcement. Built carding and chargeback defense using fingerprint analysis, GEO/BIN mismatches and low-odds betting patterns, keeping false positives within 0.5-1%.

Analyzed raw PPC traffic logs to detect anomalies, filtering click farms and incentivized traffic before advertiser billing. Identified and blocked fraud rings totaling more than $150,000, protecting advertiser budgets and affiliate reputation. Documented investigations for payout disputes and helped account managers secure full or partial payouts to webmasters and media buyers totaling over $35,000. Built data-sharing processes, reducing response time to low-quality traffic spikes by 23%.