Home Loans Loss Forecasting Analytics, Senior Data Scientist

We are looking for a Senior Data Scientist to join SoFi’s Secured Lending Team, with a focus on Home Lending risk analytics, loss forecasting, and portfolio performance monitoring. This role will support home lending products including first mortgages, jumbo loans, closed-end seconds, and HELOCs, with a strong emphasis on delinquency, default, cure, severity, recovery, and portfolio profitability.

The Senior Data Scientist will play a key role in building models, dashboards, and analytical frameworks that help the Secured Lending organization understand credit performance across the full residential lending lifecycle — from origination and portfolio monitoring through delinquency, default resolution, loss mitigation, and recovery.

This individual will partner closely with Credit Decision Science, Credit Risk, Finance, Capital Markets, Servicing, Collections, Loss Mitigation, Model Risk, and Data Engineering to support data-driven decision-making across Home Lending.

By joining SoFi, you’ll become part of a forward-thinking company that is transforming financial services for the better. We offer the excitement of a rapidly growing company with the stability of an industry-leading leadership team.

The Senior Data Scientist will help SoFi strengthen Home Lending risk analytics, forecasting, and portfolio management by:

• Developing quantitative and machine learning models to forecast losses across mortgage and home equity portfolios, including first lien, jumbo, HELOC, and closed-end second-lien products.
• Building and maintaining CECL, loss forecasting, and portfolio performance models with a focus on delinquency roll rates, default probability, cure behavior, loss severity, recovery timing, prepayment behavior, and charge-off outcomes.
• Defining and maintaining portfolio performance KPIs across credit, profitability, and risk, including delinquency rates, roll rates, cure rates, loss rates, severity, prepayment speeds, early payment defaults, repurchase risk, defect rates, and recovery performance.
• Performing cohort, vintage, and segmentation analysis by credit score, LTV/CLTV, DTI, lien position, documentation type, occupancy, channel, state/metro, property type, investor, and product type.
• Analyzing borrower behavior and identifying key risk drivers across stages of credit performance, including current status, early delinquency, late-stage delinquency, default, liquidation, foreclosure, recovery, and redefault.
• Building roll-rate models, delinquency migration analytics, cure models, default models, recovery models, and loss severity frameworks for secured lending portfolios.
• Supporting collections, loss mitigation, and default strategy analytics, including segmentation, treatment strategy measurement, liquidation waterfalls, cure versus liquidation outcomes, modification performance, and recovery optimization.
• Developing analytics that evaluate resolution pathways, including cure, modification, repayment plan, foreclosure, liquidation, REO, charge-off, and expected recovery cash flows.
• Building and maintaining executive dashboards and automated reporting that clearly explain what changed, why it changed, and what actions should be considered next.
• Partnering with Data Engineering to define data requirements, improve data quality, create new data sources, and build summarized analytical tables that support scalable reporting, monitoring, and modeling.
• Aggregating and synthesizing datasets from multiple environments, including origination data, servicing systems, collections data, collateral data, bureau data, investor/product data, and external housing market data such as HPI.
• Performing sensitivity, scenario, and stress analysis tied to home price movements, interest rates, unemployment, credit mix, prepayment behavior, and broader economic conditions.
• Monitoring model and portfolio performance through back-testing, forecast-to-actual tracking, population stability, segmentation diagnostics, drift monitoring, and periodic recalibration.
• Preparing clear, audit-ready documentation for models, assumptions, dashboards, data sources, business logic, reporting definitions, and governance routines.
• Partnering with Credit Decision Science and other cross-functional stakeholders to develop roll-rate models, collections analytics, loss forecasting enhancements, and portfolio risk insights.
• Translating complex analysis into concise, executive-ready recommendations for Credit Risk, Finance, Capital Markets, Accounting, Model Risk, and Secured Lending leadership.

• 5+ years of experience in data science, statistical modeling, credit risk analytics, loss forecasting, portfolio analytics, or a related quantitative role.
• Master’s or PhD in Statistics, Mathematics, Economics, Engineering, Computer Science, Operations Research, Finance, or another quantitative field; equivalent practical experience will also be considered.
• Strong proficiency in Python and SQL, with experience building repeatable analytical pipelines, model monitoring routines, and automated reporting.
• Experience with data visualization and dashboarding tools such as Tableau, Looker, Power BI, or similar platforms.
• Demonstrated experience with credit risk modeling, loss forecasting, CECL, roll-rate modeling, delinquency/default modeling, recovery modeling, or portfolio performance analytics.
• Hands-on experience with mortgage or secured lending data, including first liens, jumbo loans, HELOCs, closed-end seconds, or other collateral-backed products.
• Strong understanding of mortgage credit risk drivers, including FICO, LTV/CLTV, DTI, lien position, occupancy, documentation type, channel, geography, property type, investor/product, collateral value, and HPI.
• Experience analyzing delinquent, non-performing, or defaulted loan portfolios, including roll rates, cure rates, charge-offs, recoveries, redefault behavior, and severity.
• Familiarity with statistical and machine learning methods such as regression, survival analysis, time-series modeling, Markov/state transition models, gradient boosting, random forests, clustering, and model calibration.
• Strong analytical communication skills, with the ability to explain model outputs, portfolio trends, and risk drivers to both technical and non-technical audiences.
• Ability to operate in a governed risk management environment with attention to auditability, documentation, controls, and model risk expectations.