Research

My research focuses on how to design, build, evaluate, and govern AI systems that influence real-world decisions and outcomes.

Unlike much of machine learning research, which focuses on prediction accuracy, my work focuses on end-to-end systems—from problem definition to deployment to impact evaluation.

AI for Decision-Making Systems

Designing machine learning systems that support human decisions in complex policy environments.

Evaluation of AI Systems

Developing methods to rigorously evaluate AI systems in real-world settings, including randomized controlled trials and downstream outcome measurement.

Human-AI Collaboration

Understanding how humans interact with AI systems in practice, including interpretability, usability, and workflow design.

Responsible and Accountable AI

Moving beyond principles to implementation through fairness, governance, procurement, and accountability mechanisms.

Research Philosophy

Prediction alone is not useful. Impact comes from integrating prediction with action, systems, and evaluation.