Curriculum vitae
Education, research, presentations, and service.
Education
University of California, Los Angeles
Ph.D. in Biostatistics, 2022–2027 (expected June 2027). GPA 4.0/4.0.
University of California, Los Angeles
M.S. in Biostatistics, 2020–2022. GPA 4.0/4.0.
The Hong Kong Polytechnic University
B.B.A. (Hons) in Financial Services, minor in Applied Mathematics, 2015–2019. GPA 3.6/4.0.
Research interests
Prediction-powered and semi-supervised inference with machine learning and large language model predictions; causal inference for time-to-event outcomes; survival prediction and model evaluation.
Research experience
Ph.D. research, Department of Biostatistics, UCLA
- Prediction-powered Cox regression with large language model predictions. A semi-supervised Cox regression method that augments a small labeled sample with predictions, including those from large language models, for a larger unlabeled sample. Under independent censoring, its asymptotic variance is no larger than that of the labeled-only estimator, whether or not the Cox model holds. Applied to NHANES mortality and NACC dementia progression. Manuscript in preparation; Python package in development.
- Predictive performance evaluation for survival and competing risks. Time-restricted generalized R² and pseudo-R² measures for right-censored competing-risk prediction (Biometrics, in press). A tutorial on survival prediction performance measures and the R package PAmeasures (submitted to Statistics in Medicine).
- Instrumental variable estimation for censored outcomes. A data-adaptive censored two-stage least squares estimator for semiparametric accelerated failure time models (Statistics in Biosciences, 2026).
- Historical-control matching for clinical trials. A matching framework that combines propensity scores from structured covariates with large language model embeddings of clinical notes.
Epidemiological modeling and predictive analysis, The Hong Kong Polytechnic University
- Developed an interpretable XGBoost model for antituberculosis drug-induced liver injury, with an R Shiny application for clinical risk prediction.
- Built SEIR models of inter-city SARS-CoV-2 transmission and estimated inpatient and ICU bed shortages.
- Modeled mortality, long ICU stays, and length of hospital stay for COVID-19 patients.
Software
- PAmeasures (author and maintainer). Prediction performance measures for continuous and survival outcomes, including right-censored data, competing risks, and case-cohort and nested case-control designs. CRAN GitHub
- cTSLS. Censored two-stage least squares for semiparametric AFT models, with a web application. GitHub Web app
Presentations
- Time-dependent pseudo-R² for assessing predictive performance in competing risks data. Contributed talk, Joint Statistical Meetings, Boston, MA, August 2026.
- Two-stage least squares instrumental variable estimation for semiparametric accelerated failure time models with right-censored data. Contributed talk, ENAR Spring Meeting, Indianapolis, IN, March 2026.
- Two-stage least squares instrumental variable estimation for semiparametric accelerated failure time models with right-censored data. Invited seminar, Department of Biostatistics, City University of Hong Kong, September 2025.
Professional service
- Session chair, “Approaches for Measurement Error and Missing Data,” Joint Statistical Meetings, Boston, MA, August 2026.
Statistical consulting and teaching
Statistician, UCLA Jonsson Comprehensive Cancer Center
Member of the Biostatistics, Analytical Support & Evaluation (BASE) Unit: study design, analysis planning, power and sample size calculations, and statistical sections of NIH grant applications.
Statistician, UCLA COPD Program
Statistical modeling, visualization, and interpretation for clinical research in chronic obstructive pulmonary disease.
Teaching Assistant, UCLA
BIOSTAT 410 (Clinical Trials), 250A/250C (Linear and Multivariate Models), 215 (Survival Analysis), 202A/202B (Mathematical Statistics), and 200C (Generalized Linear Models).
Teaching Assistant, Emory University
Summer Institute in Statistics and Modeling in Infectious Diseases (SISMID).
Skills and interests
Programming: R, Julia, Python, Stan, SAS, LaTeX.
Interests: Tennis, badminton, table tennis, photography.