Causal Inference · Representation Learning · World Models
I'm a Ph.D. candidate in Statistics at Boston University and an AI Research Scientist Intern at Unity. My work connects causal inference, statistical machine learning, and representation learning, from developing methods to building research software.
Previously, I worked as an Applied Scientist at Amazon, a Data Scientist at Riot Games, and a Statistical Innovation and Data Science Co-op at Moderna.
- World models at Unity: task-relevant representations and automatic selection of the latent dimension for planning, building on SCALE.
- Causal inference at Boston University: high-dimensional mediation, nonlinear representation learning, latent factor models, and causal imitation learning under measurement error. I work on methodology, theoretical analysis, and simulation studies.
Debiased causal mediation analysis in ultra-high-dimensional settings with interaction effects. The R research code includes simulations, real-data analysis, and method comparisons.
Nonlinear representation learning for high-dimensional causal mediation. The Python implementation includes encoder training, cross-fitted effect estimation, and simulation evaluation.
Open-source contributions to a benchmark for AI agents working on scientific workflows. I contributed a statistics task on high-dimensional mediation debiasing.
Upstream repository · Merged contribution
For publications and research updates, see my Google Scholar.