Principal Data Scientist (Remote)

UnitedHealth Group•Salt Lake City, UT

8h•$112,700 - $193,200•Remote

About The Position

Optum is seeking a Principal Data Scientist to build and maintain predictive credit-risk models, ideally for commercial line of credit products. The role develops models for application and behavioral scoring, probability of default, loss given default, early risk detection, underwriting, fraud detection, credit line management, and collections strategies. It also involves creating advanced features from alternative data sources and evaluating data vendors to improve model performance. The Principal Data Scientist owns models from prototype through production in partnership with Data Engineering and Product teams, ensuring models are interpretable and compliant with FCRA, ECOA, and fair lending standards. The role includes model monitoring, validation, reporting, back testing, and recalibration for changes in economic conditions or borrower behavior. The position collaborates with Risk, Product, Compliance, and Finance teams to incorporate model outputs into strategy and communicate findings to technical and non-technical stakeholders. This role can be performed remotely from anywhere within the United States. Remote employees must follow UnitedHealth Group’s Telecommuter Policy.

Requirements

Master’s degree in a quantitative discipline such as Statistics, Mathematics, Data Science, Economics, or Computer Science.
5+ years of experience working with Python or R and SQL.
5+ years of experience working with lending lifecycle and credit risk KPIs, such as delinquency, charge-offs, approval rates, ROI, and CAC/payback.
4+ years of experience in credit risk modeling, ideally at a fintech, neobank, or consumer lending startup for commercial line of credit products.
4+ years of experience with traditional statistical techniques, such as logistic regression and scorecards, and modern machine learning.

Nice To Haves

Familiarity with cloud environments such as AWS, GCP, or Snowflake.
Experience with real-time underwriting or decisioning APIs.
Experience in embedded finance, BNPL, gig economy lending, or small business credit.
Familiarity with explainability frameworks such as SHAP or LIME.
Exposure to model governance or regulatory reviews.
Comfort working in a dynamic, test-and-learn environment with incomplete data.
Solid communication skills and attention to model transparency, fairness, and compliance.

Responsibilities

Develop and maintain predictive models for application and behavioral scoring, probability of default, loss given default, and early risk detection.
Design models for underwriting, fraud detection, credit line management, and collections strategies.
Create advanced features using alternative data sources, such as bank transaction data, health care claims data, cash flow, and open banking.
Evaluate new data vendors and enrich internal datasets to improve model performance.
Take models from prototype to production in partnership with Data Engineering and Product teams.
Build interpretable and compliant models aligned with FCRA, ECOA, and fair lending standards.
Own model monitoring, validation, reporting, and back testing, and recalibrate models for shifts in economic conditions or borrower behavior.
Collaborate with Risk, Product, Compliance, and Finance teams to integrate model outputs into strategy.
Communicate insights clearly to technical and non-technical stakeholders.