What are the research methods that offer the most useful data to policymakers in this rapidly changing landscape? Find out in #EvidenceInsight, a new video series from Mathematica Policy Research.
- Natural Language Processing
- Advanced Data Analytics
- Monte Carlo Simulation Design
- High-Performance Technical Computing
- Data Visualization and Presentation
- Medicare and Medicaid Data
- Health Care Bundled Payment Programs
- Supplemental Nutrition Assistance Program
- Medicaid and CHIP
- Nutrition and Food Assistance Programs
Phillip Killewald researches and implements state-of-the-art quantitative analysis techniques to help address complex policy evaluation and program improvement challenges for governments and foundations.
Killewald is a leader and key developer for many of Mathematica’s advanced analytics, machine-learning, and natural language processing offerings. He leads a team developing and testing deep neural network models for classification and topic extraction from survey data and interview transcripts for various federal government clients. As the leader of ad hoc tasks on retrospective bundled payment monitoring and implementation projects for the Center for Medicare & Medicaid Innovation, Killewald has written specifications and implemented programs for calculating bundled payment reconciliation. He also designs and implements software to run Bayesian models on large program evaluation projects for the Centers for Medicare & Medicaid Services and develops simulation models to study policy effects on Supplemental Nutrition Assistance Program participation for the Food and Nutrition Service.
Killewald joined Mathematica in 2015 from the Massachusetts Institute of Technology Lincoln Laboratory, where he was a technical staff member in the Systems and Analysis Group. In that role, he led multiple studies for the U.S. government, designing large-scale simulations of physical systems to understand key performance variables and distilling and presenting complex results to mixed audiences of technical and non-technical experts and decision makers. Before his work at the Lincoln Laboratory, Killewald conducted dissertation research at CERN, the European Organization for Nuclear Research in Geneva, Switzerland. During that time, he developed numerical methods to solve complex analytic problems and used computing tools to enable processing of immense data sets. In addition, he implemented supervised machine learning techniques to classify data-rich events. Killewald holds a Ph.D. in physics from Ohio State University.