Develop predictive models/software for commercial and financial market use. Build/deploy full-stack computational tools in cloud/container/k8s context. Design cloud-based maintanable model training/deployment paradigms. Evangelize on best practices at the intersection of predictive modeling and software development.
Created new upper-level Options/Futures course. Taught both theoretical and applied approaches to derivatives modeling.
Developed financial stress test models for DFAST, CCAR, and CECL uses, specifically commercial default probability models. Evangelized version control and coding best practices within company.
Led development team creating interactive data visualization dashboard and productivity application for federal agency department developed in Django/Python, PostgreSQL, JavaScript, and HTML. Led development team creating data dashboard for federal agency HR department developed in MS .NET C#, SQL Server, JavaScript, and HTML. Advised on econometric time series and statistical approaches to measuring return on investment for training courses. Taught team members how to develop in Python, JavaScript, HTML, and C# and how to use modern software development practices such as version control, test-driven development, and bug/issue reporting. Developed and deployed project-wide version control and issue tracking systems. Responsible for delivery and deployment of applications used directly by agency personnel making workforce decisions. Independently developed workforce modeling command line and web tool for simulating workforce events such as retirement, plant closings, with business rules input by user, written in Python.
Created new Masters-level Public Finance course. Taught both structural macroeconomic modeling approaches and time series econometric approaches to modeling the impact of fiscal shocks.
Developed complex SAS programs to import and analyze Medicare, Medicaid, and commercial health care claims (>10 GB) across multiple years' databases to produce hospital and patient level outcomes. Entrusted with automated system utilizing Perl and SAS to estimate hospital level health outcome point estimates and bootstrapped confidence intervals achieved by customized parallel processing and non-linear modelling. Introduced company-wide git source code management to Research Programming Division, expediting code review, and formalized version control. Independently developed and taught R software training courses to increase R programming support at company. Provided crucial Linux support to programming team, writing several bash and Python scripts to aid transition from MS Windows to Linux environment, in addition to monitoring Linux system performance. Reviewed SQL, R, MATLAB, and Stata code of incoming job candidates.
Contributed time series analysis functionality to the statsmodels Python module. Created the SVAR module, along with unit-root tests and impulse response function options. Contributions funded by Google Summer of Code.
Constructed and analyzed internal data sets from a variety of commercial banks involved in the sub-prime mortgage crisis. Created data used for quantities, graphs, and charts used in official report requested by President and US Congress.
In 2010, wrote SAS macros for use by the Poverty Branch of the Housing and Household Economic Statistics Division in order to run descriptive statistic queries on Current Population Survey (CPS) data going back to 1963. Worked with internal ACS and CPS data to produce statistics for the Income, Poverty and Health Insurance Coverage Reports.
In 2009, made extensive use of internal American Community Survey (ACS) data to assess the quality of new health insurance coverage data. Worked with internal ACS and CPS data to produce statistics for the Income, Poverty and Health Insurance Coverage Reports.
Dissertation: A Computational Approach to Affine Models of the Term Structure
Core expertise in time series, panel, non-linear, and unobserved component models in commercial/financial contexts.
Full-stack software development using test-driven development, continuous integration, and container orchestration.
Passion for teaching best practices for software development, production predictive modeling, version control (git), and how to merge the worlds of software development and predictive modeling.