Principal/Senior Research Scientist, Causal and Explainable Credit Risk
About the Role
This role leads the causal inference and explainability agenda within a newly formed research group building the next generation of consumer credit scoring from the ground up. You will bring deep credit-risk domain expertise to ensure models can answer counterfactual questions and produce defensible, regulator-ready explanations. The work sits at the intersection of rigorous academic research and direct real-world impact on consumer lending.
What You'll Do
Lead research on causal inference and counterfactual explanations, including deriving actionable reason codes from representation-learning models and estimating treatment effects of consumer actions on future risk.
Design comprehensive evaluation frameworks covering discrimination, calibration, economic-regime stability, fairness analysis, and reason-code accuracy.
Build and own credit-risk modeling components including default prediction, recovery and loss-given-default models, and account-level hazard models.
Translate regulatory and model-governance requirements into concrete architectural and training constraints for foundation model work.
Author methodology documentation for lender model risk teams and regulators, and present findings to those audiences directly.
What We're Looking For
10 or more years of experience building credit-risk models in production at a bank, card issuer, or advanced fintech lender.
Hands-on applied causal inference experience: double ML, uplift modeling, CATE estimation, instrumental variables, staggered difference-in-differences, or synthetic control methods on real business decisions.
Direct experience designing explainability for regulated credit decisions, including adverse action reason codes, SHAP-based attribution, and counterfactual explanation methods.
Working knowledge of FCRA, ECOA, Regulation B, and model risk management frameworks as applied to credit modeling.
Experience building default, recovery, LGD, or account-level hazard and survival models.
Ability to articulate identification assumptions in causal models and diagnose where they break down.
Experience applying transformer or sequence deep learning architectures (in Python and PyTorch) to credit or transaction data.
A quantitative PhD in mathematics, economics, statistics, or a related empirical-methods field is preferred.
Experience presenting methodology and research findings to regulatory, central-bank, or academic audiences is a strong plus.
Location
On-site. Primary location is San Francisco, CA. New York, NY and Washington, DC are also accepted locations.