Research Background

My research develops statistical methods to optimize decision making in dynamic settings, especially in the context of digital marketing, mobile health technology, and/or customer-adaptive learning algorithms, such as reinforcement learning methods. Past projects include

• Meta-analytic methods for online experimentation (in collaboration with Zillow Group)
• Count-data forecasting models for customer analysis
• Robust causal inference methods for complex dynamic experiments
• A contextual bandit algorithm for optimizing treatment decisions in longitudinal settings (e.g., when to send vs. not send reminders to patients/customers)

In my ongoing work, I am developing machine learning methods to accelerate the estimation of dynamic discrete choice models—a family of models that is often used to describe consumers’ decisions among competing alternatives over time. Another project develops robust causal inference methods for estimating moderation effects in dynamic experiments, allowing for complex patterns of cross-participant dependence.

For the most up-to-date information, see Recent Updates below or download my CV.

Academic/Professional Timeline

Recent Updates

• January 2026. Forecasting Count Data with Varying Dispersion: A Latent-Variable Approach was accepted at the Journal of Forecasting.
• November 2025. I posted a preprint of Robust Bayesian Inference of Causal Effects via Randomization Distributions to arXiv.
• October 2025. I posted a preprint of Stable Central Limit Theorems for Discrete-time Lag Martingale Difference Arrays to arXiv and submitted it for consideration at the Journal of Applied Probability.
• August 2025. I successfully defended my dissertation at the University of Michigan and started my postdoc.
• January 2025. I recently accepted a position as a postdoctoral researcher at the Johns Hopkins Carey Business School under the supervision of Professor Michael Keane (beginning August 2025).