I am a Research Data Scientist at Google.
I obtained my Ph.D. in Statistical Science from UC Santa Cruz, advised by Dr. Athanasios Kottas. My dissertation research centered on Bayesian nonparametric modeling techniques, where I developed a toolbox for ordinal regression, featuring flexible and efficient models tailored to various settings. Before joining Google, I was a postdoctoral scholar at UC San Francisco.
I enjoy solving real-world problems and find fulfillment in applying statistical innovation to drive positive change. Feel free to reach out.
🔥 News
- Jan 2025: 🎉 Our paper received the student paper award from the Risk Analysis Section of ASA. [Paper]
- Oct 2024: 🎉 Our paper is accepted by Statistics and Computing. [Paper] [Code]
- May 2024: 🎉 I successfully defended my Ph.D. dissertation in Statistical Science at UC Santa Cruz. [Slides]
🚀 Project Highlights
Structured Mixture of Continuation-ratio Logits Models for Ordinal Regression
We proposed a unified toolbox for ordinal regression by directly modeling the discrete response distribution. The virtues of the proposed models rely on the following key ingredients:
- Continuation-ratio logits representation;
- Pólya-Gamma data augmentation technique;
- Logit stick-breaking process prior.
Flexible Bayesian Modeling for Longitudinal Binary and Ordinal Responses
We developed a modeling framework for the dynamic evolution of ordinal responses from longitudinal studies. The key features of the proposed model are:
- Flexible structure for the mean and covariance;
- Unified toolbox for balanced/unbalanced longitudinal studies with binary/ordinal outcomes;
- Computationally effcient posterior simulation method.