Welcome! I’m Kai, a PhD student in the Operations, Information and Decisions Department at The Wharton School of the University of Pennsylvania, and an affiliate of the Centre for Causal Inference at Penn Biostat. I am very fortunate to be advised by Dean Knox.
I grew up in Birmingham, UK (but support Arsenal FC). Before joining Penn, I earned a master’s degree in mathematics from Imperial College London, including a year studying abroad at the École polytechnique fédérale de Lausanne in Switzerland.
When not in the office, you’ll usually find me at the cinema, playing badminton, or reading and watching dystopian science fiction.
Research interests
I study causal inference in theory and application. My research broadly follows two streams: (i) developing partial identification strategies that permit causal inference for social scientific questions in challenging data environments, and (ii) applying modern causal inference techniques to policy evaluation and fairness in social and institutional settings such as police enforcement and education.
News
Isaac Newton Institute poster presentation
Poster. Using Causal Inference to Unmask Racial Discrimination in Traffic Enforcement via Proxies at the Foundations of Causal Inference Workshop, Isaac Newton Institute.
University of Washington Causal Reading Group talk
Talk. Using Causal Inference to Unmask Racial Discrimination in Traffic Enforcement via Proxies at the University of Washington Causal Reading Group.
Data Science Frontiers workshop announcement
Workshop. Dean Knox and I will deliver a talk and workshop on applications of the autobounds software at Data Science Frontiers (Society and Politics) in New York.
IC2S2 2025 parallel talk accepted
Parallel Talk. Using causal inference to unmask racial discrimination in traffic enforcement via proxies has been accepted for IC2S2 2025 in Norrkoping, Sweden.
Poster Presentation. Using causal inference to unmask racial discrimination in traffic enforcement via proxies has been accepted for ACIC 2025 in Detroit.
Parallel Talk. Quantifying causal effects via temporal regression discontinuity designs with time-varying effects has been accepted for a 15-minute presentation at Joint Statistical Meetings 2025 in Nashville.