Manager, Data Science - US Remote

MoneyGram•New York, NY

9d•$150,000 - $200,000•Remote

About The Position

The role is responsible for designing, developing, and deploying production fraud detection models that score transactions in real time. This individual will co-own the end-to-end data science roadmap for transaction fraud and risk, build and maintain feature pipelines leveraging transaction data, device signals, behavioral patterns, and identity attributes, and lead the transition from rules-based fraud detection to a model-first decisioning architecture. They will design the interaction between models and rules, implement champion/challenger frameworks to optimize performance, and create monitoring systems for model drift, feature distribution shifts, and rule effectiveness. Additional responsibilities include generating explainability outputs for every model decision, mentoring a small team of data scientists while remaining hands-on, partnering with Risk Intelligence to align strategies with business objectives, and presenting performance analysis and strategic recommendations to leadership.

Requirements

7+ years of progressive experience in machine learning, data science, or quantitative risk
4+ years building production ML models in fraud, risk, payments, or financial services
3+ years working with rule-based fraud detection systems, including rule design, tuning, and performance optimization
2+ years leading or mentoring data scientists or analysts in a technical capacity
Demonstrated track record deploying and maintaining models in real-time production systems
Expert-level proficiency with gradient boosting frameworks (XGBoost, LightGBM, CatBoost)
Strong experience with rule engines and decision management systems
Advanced feature engineering for transactional and behavioral data
Production ML deployment including model serialization, API integration, and latency optimization
Advanced SQL for large-scale data manipulation (BigQuery, Snowflake, or similar)
Python fluency: pandas, NumPy, scikit-learn, and model deployment frameworks
Experience with model monitoring, drift detection, and automated retraining pipelines
Understanding of model explainability techniques (SHAP values, feature importance, gain importance)
Strong understanding of fraud patterns: account takeover, identity fraud, transaction fraud, or similar
Experience designing hybrid systems where models and rules work together effectively
Strong grasp of rule lifecycle management: creation, testing, deployment, monitoring, and retirement
Familiarity with identity and risk signals (device fingerprinting, phone/email intelligence, velocity patterns)
Experience balancing approval rates against fraud losses—you understand the business trade-offs

Nice To Haves

Experience with decisioning platforms (Oscilar, Datavisor, Actimize, or similar)
Background in money transfer, remittance, or cross-border payments
Experience leading organizations through transitions from rules-heavy to model-first fraud detection

Responsibilities

Design, develop, and deploy production fraud detection models that score transactions in real-time
Co-own the end-to-end data science roadmap for transaction fraud and risk
Build and maintain feature pipelines using transaction data, device signals, behavioral patterns, and identity attributes
Lead the transition from rules-based fraud detection to model-first decisioning architecture
Design the interaction between models and rules; determining when models make primary decisions versus rules
Implement champion/challenger frameworks to continuously test and improve both model and rule performance
Create monitoring systems for model drift, feature distribution shifts, rule effectiveness, and overall system performance
Generate reason codes and explainability outputs for every model decision
Mentor and lead a small team of data scientists while remaining hands-on with development
Partner with Risk Intelligence team to align model and rule strategies with business objectives
Present performance analysis, trade-off recommendations, and strategic roadmaps to leadership