I lead the Causal Evidence and Decisions Studio (CEADS) at Yale, where we develop machine learning–aided causal inference methods for high-stakes decision-making. Our work sits at the intersection of nonparametric and semiparametric statistics, interpretable ML, and data science—grounded in first-principles thinking and process-driven science.
We partner with domain experts across a vast spectrum of critical challenges—from estimating disease prevalence and understanding deforestation in Madagascar, to quantifying the socio-economic burden of health shocks. Our collaboration partners span North America, Europe, Africa, and Asia, working in public health, medicine, environmental science, ecology, and supply chain management.
Our tools are flexible, robust, and trustworthy, designed for scenarios where standard causal assumptions are challenged. We commit to delivering solutions that are accurate, trustworthy, and domain-conscious.
Combining experimental and observational data, partial identification, integrating disparate outcome measures, and characterizing underrepresented populations.
Matching methods (MALTS, VIM, AME), the Rashomon effect, distributional treatment effects, and regularized extrapolation.
Validating and benchmarking causal methods through controllable generative sandboxes and power analysis for trial design.
Causal relational learning on structured data, graph ML estimators for network effects, and transporting effects across social networks.
Quantifying heterogeneous socio-economic impacts of health shocks, including breast cancer income loss and lasting costs of illness.
Causal analysis of deforestation drivers in Madagascar and the relationship between vanilla farming and wildlife hunting pressures.
Developing frameworks for ensuring equitable data representation in public health data science and evidence-based policymaking.
Master the art of algorithmic thinking, prediction modeling, and data visualization within the context of public health. Learn to translate 'messy' real-world problems into precise mathematical solutions using Python. Course website →
An overview of recent research in interpretable machine learning, covering fundamental principles and hands-on activities with real-world data for effective analysis and responsible decision-making. Tutorial site →
I am always looking for motivated students interested in causal inference, machine learning, and public health. If you're interested in joining CEADS as a PhD student or postdoc, please reach out with your CV and a brief description of your research interests.
Members and affiliates of the Causal Evidence and Decisions Studio.
