Senior Data Scientist

You turn data into reliable, well-governed models and translate their outputs into clear recommendations for the partners who own credit, fraud, marketing, and collections strategy.\n\nKey Responsibilities\nDevelop, validate, and maintain credit risk models for underwriting and portfolio risk decisioning (scorecards, PD/loss and other supervised ML and statistical models).\nBuild and support models across additional lending decisioning use cases as prioritized by business partners, including fraud risk, marketing/propensity, customer lifecycle, collections prioritization, account management, and pricing/line/offer optimization.\nOwn the hands-on model lifecycle for your work: problem framing, exploratory analysis, feature engineering, model development, validation-ready documentation, and post-deployment monitoring.\nEngineer features and evaluate alternative and third-party data sources (bureau, alternative credit, cash-flow/bank, device, and identity signals), quantifying lift, stability, cost, and compliance trade-offs.\nProduce thorough model documentation and support model risk management and governance through validation, audit, and review cycles.\nMonitor model performance, drift, and stability in production (PSI, KS, AUC/Gini decay, calibration) and diagnose degradation, recommending recalibration or rebuild when needed.\nDesign and execute champion/challenger tests, backtests, and holdout analyses to measure model impact on approvals, losses, response, and downstream portfolio performance.\nApply explainability methods (e.g., SHAP, reason-code generation) and build fair lending and regulatory awareness into model design and feature selection (disparate-impact considerations, adverse action/reason codes).\nPartner with data engineering and decisioning-platform teams to productionize models, evolve features and pipelines, and ensure reliable real-time and batch scoring.\nTranslate model outputs into actionable business recommendations—connecting scores and segments to the policies, offers, and operational actions owned by credit risk, fraud strategy, marketing, collections, product, and other stakeholders.\nContribute to analytical standards, reproducibility, and tooling (code review, version control, experiment tracking) and help raise the technical bar across the team.\nHelp shape and refine the model roadmap by surfacing opportunities, sizing impact, and prioritizing work with the Director, Data Science and business partners.\n\nQualifications\n4-6 years in data science, statistical modeling, credit risk analytics, or a closely related quantitative field, with hands-on development of models used in real decisioning.\nStrong, current hands-on Python and SQL skills used to build, evaluate, and ship models.\nCommand of supervised machine-learning and statistical methods (e.g., gradient boosting, logistic/regularized regression, tree-based models, validation and feature-selection techniques).\nDemonstrated experience across the model lifecycle: feature engineering, model development, validation-ready documentation, and monitoring of drift and performance.\nExperience with credit risk or consumer lending models preferred (scorecards, PD/loss, or related decisioning models).\nAbility to communicate complex modeling concepts, assumptions, and trade-offs clearly to both technical and non-technical stakeholders, and to translate results into recommendations.\nBachelor's degree in a quantitative field (Statistics, Computer Science, Mathematics, Economics, Engineering, or related); advanced degree a plus.\n\nPreferred Qualifications\nConsumer lending experience including familiarity with US consumer lending regulations and risk management practices.\nFamiliarity with regulatory and model risk expectations relevant to lending models (e.g., FCRA, ECOA/Reg B fair lending, SR 11-7-style model risk management).\nExperience building or supporting models beyond credit risk, such as fraud, marketing/propensity, collections, customer lifecycle, or pricing/line optimization.\nHands-on experience evaluating alternative and third-party data and integrating models into a real-time decisioning platform.\nFamiliarity with explainability methods (SHAP, reason-code generation) and fair-lending/disparate-impact testing.\nExperience with modern ML/MLOps tooling and reproducible, version-controlled modeling workflows.