About
I'm a research scientist focused on measurement and evaluation — on whether complex sociotechnical systems are actually helping the people who use them. In online communities, that meant studying belonging, connection, and the health of social spaces. In AI, it means evaluation: is a model genuinely helpful? The surface looks different; the core challenge is the same.
The methods span causal inference, psychometrics, Bayesian modeling, and human evaluation design. The through-line is a conviction that good measurement is itself a contribution — and that getting it wrong doesn't just produce bad numbers, it produces bad decisions.
Career
OpenAI
Member of Technical Staff, Research & Product
Sole data scientist in OpenAI’s Research organization; built evaluation methodology and human-data infrastructure used across frontier model development.
Head of Research Science · Staff Research Scientist
Led community governance and quality measurement research; mixed-methods studies of communities and founders informed company-wide changes to product and organization.
Twitch / Amazon
Head of Science, Community Health · Senior Research Scientist
Founded and led the 5-person research and data science team for Twitch’s Community Health organization; built the platform’s primary harm-prevalence measurement pipeline.
Stanford University
PhD, Computer Science (HCI) · Advisor: Jeffrey Heer
PARC
Research Assistant · Peter Pirolli, Ed H. Chi
Education
Ph.D. in Computer Science (HCI) from Stanford, advised by Jeffrey Heer. M.A. in Philosophy and B.S. in Mathematics, also from Stanford. 1,700+ citations across CHI, CSCW, ICWSM, EuroVis, WSDM, UIST, and journal venues.
For detailed work history, see my full CV.