Data Scientist (Remote Eligible)

Mathematica applies expertise at the intersection of data, methods, policy, and practice to improve well-being around the world. We collaborate closely with public- and private-sector partners to translate big questions into deep insights that improve programs, refine strategies, and enhance understanding. Our work yields actionable information to guide decisions in wide-ranging policy areas, from health, education, early childhood, and family support to nutrition, employment, disability, and international development.

About the role

We are looking for a Data Scientist who will derive meaning from data through the creation and deployment of data-driven approaches to solve problems and answer important policy questions for clients. A Data Scientist owns data processing and analysis tasks and supports more senior level data science staff in implementing statistical, machine learning, generative AI, and other data science methods for use in research reports, internal systems, or client systems. This position will work remotely or flexibly in one of our office locations.

Responsibilities

• Conduct causal, predictive, and descriptive analyses
• Write and maintain programming systems in languages such as Python and R to build and evaluate models
• Develop reliable data pipelines to obtain, combine, and transform datasets on cloud, internal, and client servers
• Communicate technical results to diverse stakeholders including clients and cross-functional teams
• Develop and maintain technical and methodological documentation
• Co-develop analysis plans with a senior data scientist or researcher
• Lead and manage small teams and tasks with oversight from a more senior staff member

Example Projects

• Build and evaluate generative AI tools to extract clinically important information from unstructured doctors’ notes, then use that information to construct predictive models and descriptive statistics to improve doctor decision-making and predictive accuracy.
• Evaluate and monitor the impacts of an alternative payment model for primary care in terms of care quality, cost, and health outcomes for diverse beneficiaries, using claims from thousands of primary care practices across the country. Use the same data to predict future hospital costs and behavior.
• Analyze nationwide geographic access to food retailers by integrating geospatial data on retailer locations, neighborhood demographics, demand, and social vulnerability. Apply network-based accessibility analyses to compare convenient access within and across states overall and by urbanicity and retailer type and develop interactive dashboards that help policymakers identify disparities and improve access to nutrition assistance.
• Use national survey data and grocery store purchase data to simulate realistic American diets and analyze their nutritional value. Analyze how that nutritional value compares to guidelines and what it suggests are practical, culturally aware food baskets consumers might purchase to meet the guidelines.
• Build knowledge synthesis solutions for government and foundation clients leveraging NLP and GenAI methods (knowledge graphs, Model Context Protocol, retrieval-augmented generation) to extract quantitative information (e.g., summary statistics, regression results) and contextual details (e.g., implementation specifics, focus group discussion themes) to distill large literatures into digestible datasets that support evidence-informed policymaking.
• Develop and evaluate a reproducible benchmarking pipeline to compare state healthcare spending against peer markets nationwide, harmonizing multi-source claims and Census data, applying statistical matching to select comparable regions, and normalizing spending through risk-adjusted regression models to support state rate-setting decisions.
• Build and evaluate interpretable machine learning models to predict clinical care tiers from health assessment data, supporting state healthcare program's transition to a new assessment tool. Partner with subject-matter experts to engineer clinically meaningful features from raw assessment items, and apply stratified sampling and diagnostics to deliver transparent models suited to high-stakes eligibility and reimbursement decisions.

Requirements

• Master’s degree in a technical field such as statistics, data science, data analytics, mathematics, operations research, computer science, and/or social science; equivalent years of experience can be substituted
• Demonstrated interest and/or experience using data science and/or statistics to contribute to projects with a policy/social impact in academic and/or professional settings
• Experience applying generative AI programmatically to extract insights from unstructured data, construct new features for analysis, or as a part of a larger systematic analysis
• Experience executing causal, predictive, and descriptive data science and statistics techniques including regression modeling, machine learning algorithms, network analysis, or natural language processing
• At least three years of experience performing data cleaning and analysis using programming languages such as R, Python, or Julia in the academic, extra-curricular, or professional environment
• Ability and desire to work independently and take initiative as part of an interdisciplinary team that may be geographically dispersed. This includes being able to learn from resources such as academic articles, white papers, self-guided tutorials, and package documentation and willingness to constantly learn and contribute to knowledge sharing with team members
• Experience with reproducible research principles, version control, interactive visualizations, and common packages/libraries for supporting data science work in R, Python, and/or Julia (e.g., tidyverse, data.table, R Shiny, R Markdown, pandas, polars, NumPy, scikit-learn, MLJ.jl, DataFrames.jl, and/or Makie.jl)

Qualifications

Desired but not required: experience with healthcare datasets (for example, Medicare or Medicaid claims and enrollment data), production-quality machine learning applications, cloud computing environments (AWS/Databricks/Snowflake/etc.), and algorithmic fairness and ethics.

Benefits

Mathematica offers competitive salaries and a comprehensive benefits package, as well as the advantages of being 100 percent employee owned. As an employee stock owner, you will experience financial benefits of ESOP holdings that have increased in tandem with the company’s growth and financial strength. Learn more about our benefits here.

Pay

This position offers an anticipated annual base salary range of $70,000 - $90,000.

Staff in our Data Solutions division may eventually work with some of our largest clients, including the Centers for Medicaid & Medicare Services (CMS) and other agencies. Most staff working on these contracts will be required to complete a successful background investigation including the Questionnaire for Public Trust Position SF-85. Staff that are unable to successfully undergo the background investigation may need to obtain work outside these contracts.