Functioning as Senior Scientist, I lead a 15-member cross-functional team focused on Automation and Operational RnD. Following agile methodologies and within an over-arching MLOps framework, we develop and deploy comprehensive automation architectures. These meld machine learning and other algorithm deployments with comprehensive monitoring, scenario-based forecasting, and simulation capabilities to develop complete solutions for operational task-replacement and associated R&D. My duties involve overall team leadership, mentoring, strategy development, operational planning and tracking, and C-suite communications. As a hands-on leader, I also make significant scientific SME contributions in the following areas: (1) development of ML models (xgboost, light gbm, MARS, random forests, elastic nets, statistical clustering), (2) applied statistics and demography, (linear and generalized linear models, including hierarchical models, demographic analysis and projections including mortality models, time-series modeling & forecasting, survival analysis (proportional hazards, AFT), (3) operational R&D (decision trees & probabilistic risk and decision analysis, system dynamics, computational simulation, optimization, queueing models), (4) deep-dive R&D focused on insights analysis, ongoing risk monitoring, ML model evaluations, and monitoring of bias & potential proxy discrimination in ML and other algorithmic models, and (5) behavioral science & epidemiologic research. Software utilized on a day-to-day basis comprise a modern tech-stack including: (1) scripting languages & associated IDEs: R (Rstudio, Posit Cloud), Python (Jupyter Labs, AWS Sagemaker, Sagemaker Studio), (2) GitHub for collaboration and documentation, (3) AWS for model deployment, (3) Dremio for universal data access and management using SQL, (4) mapping & GIS systems including QGIS, Arc-GIS Explorer, and Google Earth Pro, and (5) PowerBI & Markdown reports for monitoring purposes.

Functioning as Senior Scientist, I led an 11-person Operational RnD group composed of scientists, underwriters, actuaries, and MDs focused on development and deployment of solutions for automated/accelerated underwriting. Team competencies and tasked included: epidemiology and biostatistics, machine learning & data science, automated bias testing, applied demography & statistics, forecasting, computational modeling, cloud computing, actuarial science, and medicine.

I provided computational, data science, and analytics solutions and support to a variety of projects spanning the analysis of population health dynamics, probabilistic risk and decision models, process and cost offset analyses, and the generation of value associated with specific well-being and health promotion programs. To accomplish this, I have built a scaled modeling system that links microsimulation and agent-based modeling, machine learning, biostatistical analyses, and demographic projection to produce actionable insights and feed operational process models built on system dynamic approaches. To support these activities, I regularly re-structure and extract very large and complex databases using SAS and R.

I served as a member of the National Academies of Science's Panel to Review the 2010 Census from 2009-2015 and the Standing Committee on Re-engineering the 2020 Census from 2015-2017. I focused on updating of geographic data resources and use of these resources in operations involving GIS-based capabilities.

At GPS, I led a multi-dimensional team aimed at improving both improving basic research around small-area demographic forecasts and their application to real-world problems. This research resulted in 15 publications in peer-reviewed scientific journals, two Springer monographs on demographic methods, and over 50 invited presentations and conference papers, including those made to both the New Mexico State Legislature and the United States Congress. This body of research was used to leverage ~ $2 million dollars in research funding (including funds from both National Science Foundation and National Institutes of Health) and to support research teams ranging between 5 and 15 members. Demographic estimates created in this research were used to support public health planning, creation of large state-wide geospatial databases used in a variety of applications , and were utilized by decision-makers at every level of the New Mexico State Government. Accomplishing these aims involved the creation and leveraging of large-scale, big-data geospatial databases that made this research and consulting business successful.