Built a 0?1 payment-risk intelligence MVP that evaluates merchant trust before checkout, combining multiple fraud signals into an explainable risk rating, supporting evidence, and recommended user action.
Conducted iterative usability testing with 10 users, using observed decision behavior and qualitative feedback to refine warning hierarchy, evidence presentation and recommended actions across successive MVP versions.
Built a 236-example labeled evaluation benchmark spanning 96 fraudulent and 140 legitimate cases, using leakage- controlled calibration and holdout splits to systematically evaluate fraud-decisioning performance.
Instrumented the risk engine to measure precision, recall, false-positive rate, coverage, and assessment latency, using error analysis to refine thresholds, signal weighting, and insufficient-evidence handling.
Built a greenfield data-quality capability for Orbis covering 360M+ records after identifying recurring customer friction around inconsistent entity data & manual validation across AML, KYC, Sanctions & risk workflows; defined product strategy, roadmap & success metrics, aligned 2 Engineering & 6 Data teams, & increased NPS by 12 points over 6 months.
Mentored 3 Associate Product Managers on data-product strategy, governance, modeling and prioritization.
Led a build-vs-buy evaluation for data across 10M+ records, defining vendor selection criteria and the business case across data coverage, quality, and total cost; increased data coverage by 25% while reducing costs by 10% over two quarters.
Reduced Compliance Catalyst risk-assessment processing time by 88%, from 2 hours to 15 minutes, after identifying the core workflow bottleneck through user interviews, UAT and Heap analytics; redesigned the end-to-end experience and prioritized the solution with Engineering.
Improved data consistency from 67% to 82% by establishing data-format error rate as the core product-quality KPI and driving an MVP validation solution using SQL and Python; translated requirements into validation rules and workflows.
Owned an LLM-powered financial data imputation product across 10,000+ U.S.-listed entities; partnered with Data & Engineering on validation and confidence thresholds, achieving 78% prediction accuracy.
Established the Golden Data Record framework across 4 business functions, standardizing entity definitions and governance rules; used SQL/Python to validate product hypotheses and RICE to prioritize roadmap opportunities, reducing manual reconciliation effort by 14% within 3 months and supporting expansion into 2 new verticals.
I lead complex digital products from problem discovery to launch, with a focus on AI-driven workflows, financial data, risk, and operational automation. My work spans product strategy, customer research, technical scoping, cross-functional delivery, and measurable product improvement.
Built a 0?1 payment-risk intelligence MVP that evaluates merchant trust before checkout, combining multiple fraud signals into an explainable risk rating, supporting evidence, and recommended user action.
Conducted iterative usability testing with 10 users, using observed decision behavior and qualitative feedback to refine warning hierarchy, evidence presentation and recommended actions across successive MVP versions.
Built a 236-example labeled evaluation benchmark spanning 96 fraudulent and 140 legitimate cases, using leakage- controlled calibration and holdout splits to systematically evaluate fraud-decisioning performance.
Instrumented the risk engine to measure precision, recall, false-positive rate, coverage, and assessment latency, using error analysis to refine thresholds, signal weighting, and insufficient-evidence handling.
Built a greenfield data-quality capability for Orbis covering 360M+ records after identifying recurring customer friction around inconsistent entity data & manual validation across AML, KYC, Sanctions & risk workflows; defined product strategy, roadmap & success metrics, aligned 2 Engineering & 6 Data teams, & increased NPS by 12 points over 6 months.
Mentored 3 Associate Product Managers on data-product strategy, governance, modeling and prioritization.
Led a build-vs-buy evaluation for data across 10M+ records, defining vendor selection criteria and the business case across data coverage, quality, and total cost; increased data coverage by 25% while reducing costs by 10% over two quarters.
Reduced Compliance Catalyst risk-assessment processing time by 88%, from 2 hours to 15 minutes, after identifying the core workflow bottleneck through user interviews, UAT and Heap analytics; redesigned the end-to-end experience and prioritized the solution with Engineering.
Improved data consistency from 67% to 82% by establishing data-format error rate as the core product-quality KPI and driving an MVP validation solution using SQL and Python; translated requirements into validation rules and workflows.
Owned an LLM-powered financial data imputation product across 10,000+ U.S.-listed entities; partnered with Data & Engineering on validation and confidence thresholds, achieving 78% prediction accuracy.
Established the Golden Data Record framework across 4 business functions, standardizing entity definitions and governance rules; used SQL/Python to validate product hypotheses and RICE to prioritize roadmap opportunities, reducing manual reconciliation effort by 14% within 3 months and supporting expansion into 2 new verticals.
I lead complex digital products from problem discovery to launch, with a focus on AI-driven workflows, financial data, risk, and operational automation. My work spans product strategy, customer research, technical scoping, cross-functional delivery, and measurable product improvement.