Fraud Data Scientist

As a Fraud Data Scientist at Coinflow, you'll be the analytical brain behind our fraud models. You'll analyze how models and rules perform, spot where fraud is shifting, and recommend the features, signals, and strategy changes that will move the metrics. You work across card, bank, and alternative payment rails serving merchants in gaming, marketplaces, remittance, and ecommerce, where fraud patterns shift weekly.

What you'll do

• Track fraud model and rules performance over time and by merchant segment, digging into approval rates, fraud loss, false positives, and drift to explain what changed and why
• Spot emerging fraud patterns early (fraud rings, account takeover, card testing, friendly fraud, bonus abuse), size the impact, and get ahead of it before it shows up in chargebacks
• Explore transaction, device, behavioral, identity, and network data to find new signals, prototype and validate features, then give engineering a clear, evidence-backed case for what to add to the models
• Own ground truth by making sense of delayed and noisy labels from chargebacks, disputes, and manual review, and flag where bad labels are misleading the models
• Design and analyze backtests, shadow tests, and champion/challenger experiments to evaluate third-party fraud vendors against your own models on your data
• Partner with risk operations to tune thresholds, rules, and step-up authentication, recommending decision strategies per merchant segment and quantifying the tradeoff between loss and approvals
• Turn analysis into clear reporting for leadership, merchants, banks, and auditors on what drives model decisions, performance trends, and your response strategy

What you bring

• 4+ years as a data scientist or analyst working on fraud, risk, or abuse, ideally at a payments company, fintech, or marketplace
• Proven track record using analysis to measurably improve production fraud models or rules, delivering new features, better thresholds, lower loss, or higher approvals
• Strong SQL and Python skills, plus fluency with the stats and ML concepts behind fraud models (gradient-boosted trees, calibration, imbalanced data)
• Solid grasp of the fraud problem space: card-not-present fraud, account takeover, card testing, friendly fraud, chargeback reason codes, and how fraudsters adapt
• Comfort with messy, imbalanced, delayed-label data and the judgment to evaluate model performance honestly on it
• Experience designing and analyzing experiments where a false positive and a false negative cost very different amounts
• Strong feature intuition: you can look at a fraud pattern and propose the signal that would have caught it
• Ability to explain model behavior and tradeoffs to merchants, risk ops, banks, and auditors in plain language
• A bias toward action; you'd rather deliver a clear recommendation this week than a perfect analysis next quarter

Nice to have

• Familiarity with graph or network analysis for detecting linked accounts and fraud rings
• Comfort working with AI-native tooling as a daily driver, or genuine excitement to start

What we offer

• Base salary of $120,000-$180,000 USD plus equity

About Coinflow

Coinflow is a next-generation payment service provider founded in 2023 that uses stablecoins, AI-driven fraud prevention, and instant settlement to power global financial infrastructure. Since its seed round in 2024, the company has achieved 23x revenue growth and scaled to multi-billion-dollar annual transaction volume, recently announcing a $25M Series A led by Pantera Capital, CMT Digital, Coinbase Ventures, Jump Crypto, and Reciprocal Ventures. Coinflow is headquartered in Chicago, IL.