Product Manager · Payments & FinTech · AI & GenAI Products · Fraud Prevention & Risk
T-shaped Product Manager with 11 years of software delivery — 5 years owning high-impact products at the intersection of Payments, Fraud Prevention, Financial Crime, KYC/AML compliance, and Generative AI across regulated global markets. I turn complex technical and regulatory constraints into elegant, scalable products that move real business metrics.
How I Think
In payments and financial crime, the hard part isn't the feature — it's the regulation, risk, and reconciliation around it. I translate RBI, FCA, PCI DSS, and GDPR constraints into product experiences that feel effortless yet stay audit-ready.
I ship GenAI where it earns its place — agentic fraud workflows, RAG pipelines, LLM anomaly detection — always paired with verification. AI should reduce risk and manual effort, never introduce ambiguity where money moves.
Shipping is the start line. What counts is $10M+ saved, 57% fewer chargebacks, 295% fraud efficiency, 99.98% reconciliation accuracy — measurable impact delivered through data, cross-functional leadership, and relentless focus.
How I Think · AI-Native
Building great AI products in a regulated payments platform isn't "add an LLM." For every capability I ask which era should own it — explicit rules, a trained model, or an LLM reasoning over live context — then design the verification and autonomy around it.
Sanctions screening, velocity checks, eligibility and compliance logic stay explicit — rules you can prove, log, and a regulator can trust. No ambiguity where money and KYC/AML obligations matter most.
Fraud scoring, decline patterns, and mule-account signals learn from millions of real transactions — turning messy history into calibrated risk decisions with fewer false positives.
Agentic fraud decisioning on AWS Bedrock and Claude reasons over live account and transaction context via RAG — always paired with a verification step before any consequential RiskOps action.
I place each capability deliberately on the autonomy slider — automating where outcomes are verifiable and safe, escalating to a human RiskOps analyst with full context where trust matters. Earned step by step, never assumed.
Selected Work · Entain (BetMGM)
Re-architecting fraud scoring, authorization, tokenization, and validation logic end-to-end.
A regulated platform processing 500,000+ monthly transactions across the US, UK, and Europe was leaking margin to fraud and chargebacks, with authorization and validation logic that hadn't kept pace with scale.
Owned end-to-end vision and roadmap, redesigning fraud scoring models, authorization-rate workflows, tokenization pipelines, and payment validation logic — aligning Engineering, Risk Ops, Compliance, Finance, and QA.
Unifying four fraud and identity platforms behind CDD, EDD, and sanctions screening.
Fraudulent volume and identity risk needed a layered defense no single vendor could provide. I mapped the decisioning journey across onboarding, transaction, and payout to place each control where it mattered.
Integrated four platforms with CDD, EDD, and sanctions-screening workflows, wiring real-time rule engines and velocity checks into the decision flow.
GenAI fraud workflows on AWS Bedrock & Claude — RAG pipelines and LLM anomaly detection.
Manual RiskOps review was a bottleneck to real-time decisioning at scale, and static rules missed nuanced, contextual fraud signals.
Shipped AI-powered fraud decisioning: agentic workflows with RAG pipelines and LLM-based anomaly detection reasoning over live context — every consequential action gated by a verification step.
API design, webhooks, routing, failover, and real-time event processing across channels.
End-to-end PSP integrations covering API design, webhook management, payment routing logic, failover handling, and real-time event processing across all payment channels.
Thought Leadership
Founder of The AI PM (@TheAI_PM) — a free educational series teaching AI concepts, GenAI tools, agentic workflows, and AI-first product thinking to Product Managers. A 50-episode curriculum covering LLMs, RAG, AI agents, prompt engineering, automation (n8n, Make, Claude), and applied AI for PM workflows.
Career Journey
11 years of software delivery — from QA leadership to owning payments, fraud, and AI products.
Owned end-to-end vision, strategy, and roadmap for payment gateway, fraud prevention, financial crime, and KYC/AML compliance across a regulated platform processing 500K+ monthly transactions in the US, UK, and Europe. Delivered $10M+ annual savings, 57% chargeback reduction, integrated 4 fraud/identity platforms, shipped agentic AI fraud decisioning, and maintained 95% on-time delivery across 4 consecutive quarters.
Analyzed 3M+ monthly transactions with SQL and dashboards to surface fraud patterns and payment root causes — reducing payment failure rates by 18%. Ran 15+ A/B tests on fraud scoring and checkout flows, cutting false declines by 15% while maintaining PCI DSS and GDPR readiness with zero critical defects.
Led a QA team of 8 delivering end-to-end automation for a global certification platform serving 1M+ annual candidates — reducing manual regression effort by 60% across 40+ quarterly releases via Angular and C# frameworks, at 99.9% uptime. Built the systems thinking and delivery discipline that define my PM practice today.
Expertise
AWS Bedrock, Claude, OpenAI API, GitHub Copilot
n8n, Make, Zapier, No-Code AI
Sift, LexisNexis, Incode, Prove
Jira, Confluence, Redash, Looker, Postman, Figma