Fraud Data Scientist

Coinflow•Chicago, IL

1d•$120,000 - $180,000

About The Position

Coinflow is hiring a Fraud Data Scientist to analyze the performance of fraud models and rules, identify shifting fraud patterns, and recommend features, signals, and strategy changes. The role covers card, bank, and alternative payment rails and merchants in gaming, marketplaces, remittance, and ecommerce. The Fraud Data Scientist will partner with engineering, risk operations, and compliance, using data to inform actionable decisions and AI-native tools to support analysis, feature exploration, and experimentation. The position owns analysis of model performance, fraud trends, feature opportunities, label quality, experiments, rules and risk strategies, and reporting. It includes evaluating model and rule behavior by merchant segment, investigating fraud losses and false positives, validating potential model features, assessing noisy and delayed labels, and comparing third-party vendors with internal models. The role also communicates model behavior, performance, and tradeoffs to leadership, merchants, banks, and auditors. Coinflow is a global payment service provider founded in 2023 and headquartered in Chicago, Illinois. The company provides instant settlement, fraud and chargeback indemnity, global pay-ins, multi-currency FX, and unified payouts. The salary range is $120,000–$180,000 USD plus equity.

Requirements

At least 4 years as a data scientist or analyst working on fraud, risk, or abuse, ideally at a payments company, fintech, or marketplace.
Experience measurably improving production fraud models or rules through features, threshold changes, lower losses, or higher approval rates.
Strong SQL and Python skills.
Fluency in statistical and machine-learning concepts relevant to fraud models, including gradient-boosted trees, calibration, and imbalanced data.
Solid understanding of card-not-present fraud, account takeover, card testing, friendly fraud, chargeback reason codes, and how fraudsters adapt.
Comfort working with messy, imbalanced data and delayed labels, with sound judgment in evaluating model performance.
Experience designing and analyzing experiments where false positives and false negatives have different costs.
Strong feature intuition and the ability to propose signals that could detect fraud patterns.
Ability to explain model behavior and tradeoffs in plain language to merchants, risk operations, banks, and auditors.
Comfort using AI-native tools as a daily driver or genuine excitement to begin doing so.
Bias toward action and ability to deliver timely, clear recommendations.

Nice To Haves

Familiarity with graph or network analysis for detecting linked accounts and fraud rings.

Responsibilities

Track fraud model and rule performance over time and by merchant segment, including approval rates, fraud loss, false positives, and drift; explain performance changes.
Identify emerging fraud patterns, estimate their impact, and help the team address them before they result in chargebacks.
Explore transaction, device, behavioral, identity, and network data to identify, prototype, and validate potential model features; provide engineering with evidence-backed recommendations.
Assess delayed and noisy labels from chargebacks, disputes, and manual review, and identify label problems that may mislead models.
Design and analyze backtests, shadow tests, and champion/challenger experiments.
Evaluate third-party fraud vendors against internal models using Coinflow's data and recommend which to trust.
Partner with risk operations to tune thresholds, rules, and step-up authentication; recommend merchant-segment decision strategies and quantify loss-versus-approval tradeoffs.
Report on model decision drivers, performance trends, and mitigation plans to leadership, merchants, banks, and auditors.