What are the responsibilities and job description for the Director, Predictive Modeling position at National Mortgage Insurance Corporation?
POSITION SUMMARY
The Director of Predictive Modeling will enhance and validate advanced predictive models that forecast mortgage defaults, policy cancellations, loan prepayments, and claim payments. The models will inform the company’s budgeting, portfolio review, stress testing, and reserves functions. The Director is expected to take a hands-on approach to solving data and programming challenges.
Reporting directly to the Vice President of Predictive Modeling, the Director is responsible for model design, feature engineering, model training, backtesting, and model results analysis. The role requires effective cross-team collaboration between Risk, Finance, Data, and Technology.
JOB DUTIES & RESPONSIBILITIES
- Lead and manage a small team that specializes in machine learning and credit risk modeling.
- Enhance scalable, accurate, and explainable predictive models to forecast mortgage insurance claims and expenses for both existing policies and new insurance written.
- Write high-quality SQL and Python code to prototype and implement models.
- Transform multiple forecasts (including claim frequencies, severities, amortization, persistency, new insurance written, and economic scenarios) into a comprehensive cash flow engine for budgeting, portfolio review, stress testing, and reserves functions.
- Analyze and communicate model results to stakeholders, including senior management, auditors, and regulators.
- Enhance and implement time-varying features (predictors) such as marked-to-market Loan-to-Value (LTV) ratios and current property values using large, complex data sets.
- Provide expert support and guidance to management on incorporating model results into budgeting, portfolio review, stress testing, and reserves functions.
- Optimize model execution to ensure timely delivery of results from a large portfolio of in-force mortgages.
- Ensure all models comply with internal documentation standards, model governance, change management, and version control policies.
- Enhance and execute ongoing monitoring of model performance, feature drift, and target drift.
SKILLS & KNOWLEDGE
- PhD or master’s degree in Data Science, Statistics, Econometrics, or a related field.
- 10 years of experience in financial services and data science, or a related field.
- 5 years of people leadership experience.
- Strong knowledge of econometric modeling and machine learning applications (e.g., logistic regression, XGBoost, cross-validation, probability calibration).
- Demonstrated mastery of Python or R programming, with contributions to open-source libraries for statistics or machine learning.
- Experience using Python, SQL, and GitHub.
- Exceptional written, verbal, and presentation skills for communicating with stakeholders, including senior management and external auditors.
- Experience analyzing single-family loan performance data sets from Fannie Mae, Freddie Mac, or McDash.
- Experience forecasting mortgage payment statuses using a transition matrix model.
- Experience incorporating mortgage rates, regional unemployment rates, and MSA-level home price forecasts into predictive models.
- Experience developing or validating mortgage or consumer lending credit loss models.
- Familiarity with the Single-family FHFA Mortgage Analytics Platform (FMAP).
Base Salary Range: $156,800 - $200,000/year
National MI is an Equal Employment Opportunity employer and is committed to a diverse workforce. We value and actively seek to recruit, develop, and retain individuals with varied backgrounds and experiences reflecting the full diversity of the communities that we serve. It is the policy of National MI to comply with all applicable laws concerning the employment of persons with disabilities.
Salary : $156,800 - $200,000
