Director of Data Science
About the role
The Director of Data Science — Credit Risk & Decisioning will own ClarityPay's predictive modeling strategy for consumer credit. You will lead the end-to-end development of Probability of Default (PD) models, Loss Given Default (LGD) frameworks, and behavioral scoring systems that power our origination and portfolio management decisions.
Key Responsibilities
Own the full lifecycle of Probability of Default (PD) models for installment loan and BNPL originations — from feature engineering through champion/challenger deployment and ongoing monitoring
Build and maintain LGD and EAD models to support expected loss calculations and pricing optimization
Design and execute A/B experiments to continuously improve model performance against AUC, KS, and Gini benchmarks
Define and maintain decision scorecards and cutoff strategies across product tiers, balancing approval rate, risk appetite, and margin targets
Partner with Pricing to ensure PD output feeds directly into IRR-based pricing frameworks — including Purchase Price and MDR optimization for our merchant network
Build real-time model serving pipelines in collaboration with the Data Engineering team
Drive policy rule development and scorecard governance in alignment with fair lending requirements (ECOA, FCRA)
Establish performance monitoring frameworks: PSI, CSI, and vintage-level deviation tracking versus forecast
Lead model recalibration and rebuild cycles in response to portfolio drift, macro shifts, or product expansion
Produce model documentation and validation artifacts that meet institutional investor and warehouse lender standards
Interface with external model validators and auditors as the company scales its capital markets program
Leadership & Cross-Functional Impact
Hire, mentor, and grow a team of data scientists, setting standards for modeling rigor and code quality
Be a thought partner to the CRO, CFO, and Capital Markets team on risk appetite, product design, and investor reporting
Represent ClarityPay's modeling approach to warehouse lenders, ABS investors, and rating agencies during due diligence
Required
10+ years of experience in quantitative modeling, with at least 3 years focused on consumer credit risk
Deep, hands-on expertise building PD models — logistic regression, gradient boosting (XGBoost/LightGBM), survival models — in a production lending context
Strong Python (pandas, scikit-learn, statsmodels) and SQL skills; experience deploying models to production environments
Experience with installment loan, personal loan, or BNPL products strongly preferred; point-of-sale or retail credit a plus
Fluency in credit bureau data (Experian, Equifax, TransUnion) and tradeline-level feature engineering
Proven track record building models that improved loss performance or expanded approval rates at a measurable scale
Strong communication skills: ability to translate model outputs into business decisions for non-technical stakeholders
MS or PhD in Statistics, Mathematics, Economics, Computer Science, or related quantitative field (or equivalent experience)
PREFERRED
Experience at a fintech lender, BNPL company, or marketplace lender
Familiarity with CECL / IFRS 9 expected loss frameworks
Experience presenting model frameworks to institutional investors or during ABS securitization diligence
Exposure to fair lending testing (disparate impact analysis, adverse action analysis)
Prior people management experience or demonstrated mentorship of junior data scientists