Data Scientist, Fraud Risk

Role overview

A Data Scientist role embedded within a Risk team that protects co-branded credit card programs while keeping the application experience fast and welcoming. The role owns the modeling and analytics behind onboarding fraud and identity decisioning, from application submission through account opening. It partners with Fraud Strategy, Operations, Product, Engineering, Compliance, and Credit Strategy to ship production-grade controls against identity theft, synthetic identity, first-party fraud, and coordinated application abuse.

Responsibilities

- Own and continuously improve onboarding fraud decisioning across the full application journey, including identity verification, KYC controls, fraud models, policy rules, decisioning waterfalls, and manual-review strategy. - Build, validate, deploy, and monitor models that detect identity theft, synthetic identity, first-party fraud, and coordinated abuse using identity, device, behavioral, application, bureau, network, and consortium signals. - Evaluate third-party fraud and identity vendors by testing scores and attributes, measuring incremental lift, overlap, coverage, stability, latency, and cost, and recommending when to add, replace, or retire signals. - Design and analyze A/B tests, shadow tests, holdouts, and champion/challenger strategies that balance fraud losses against approval rate, false positives, verification friction, and manual-review volume. - Investigate emerging fraud patterns and decision misses, partnering with Fraud Operations to translate case-review findings into new features, rules, models, and review strategies. - Build monitoring and AI-powered workflows that surface model drift, population shifts, vendor degradation, data-quality issues, and new attack patterns, recommending adjustments for human review.

Requirements

- 5–8+ years of experience in data science, risk analytics, or a related quantitative field, ideally at a high-growth startup or fintech. - Strong Python and SQL skills, with the ability to build models, transform raw data, and create custom datasets from complex financial data. - Hands-on experience building and evaluating predictive models in production. - Demonstrated ability to communicate findings clearly to senior leadership and external partners.

Nice to have

- Experience with application or onboarding fraud: identity theft, synthetic identity, first-party fraud, application manipulation, or fraud rings. - Familiarity with KYC, CIP, identity verification, document verification, device intelligence, behavioral signals, consortium data, credit bureau data, or alternative data. - Experience evaluating and integrating third-party fraud or identity vendors and measuring incremental value. - Experience with real-time scoring, decision engines, rules platforms, APIs, or production ML systems. - Background partnering with fraud operations or investigations teams and turning case findings into scalable controls. - Familiarity with credit card underwriting, consumer lending, or regulated financial products. - Graph, anomaly-detection, or weakly supervised methods for coordinated or emerging fraud patterns.

Benefits and work setup

- Stack: Python and SQL for modeling and analysis, Snowflake for warehousing, AWS infrastructure, and dashboarding/monitoring tools for production systems. - Competitive compensation and equity packages. - Hardware budget for a configured work computer of your choice. - Flexible paid time off. - Fully covered, high-quality healthcare, including dependent coverage, with additional access to One Medical and FSA enrollment options. - 20 weeks of paid parental leave for primary caregivers and 8 weeks for all new parents. - Access to industry-leading tooling across business units.