About Me

I am a Postdoctoral Scholar in Statistics at Stanford University, where I work with Scott Linderman and David Sussillo. My research is broadly in statistical neuroscience, probabilitic machine learning, and sequence modeling. Before Stanford, I obtained a PhD in Neuroscience at Princeton University where I worked with Jonathan Pillow on a variety of topics in statistical neuroscience. During that time I completed a research internship at Facebook Reality Labs. Before that I received an MPhil in Engineering at the University of Cambridge and a B.S. in Electrical Engineering from Michigan State University.

My CV is available here . You can get in touch with me at david dot m dot zoltowski at gmail.com.

News

[09/2025]: Two papers on parallelizing MCMC and nonlinear sequence models were accepted to NeurIPS 2025!

Recent Publications

Additional Publications

Parallelizing MCMC Across the Sequence Length

David Zoltowski*, Skyler Wu*, Xavier Gonzalez, Leo Kozachkov, Scott W. Linderman

Paper Code

arXiv 2025; To appear in NeurIPS 2025

We develop methods to parallelize Gibbs, MALA, and HMC sampling across the chain length using parallel Newton methods.
Predictability Enables Parallelization of Nonlinear State Space Models

Xavier Gonzalez, Leo Kozachkov, David Zoltowski, Kenneth L. Clarkson, Scott W. Linderman

Paper

arXiv 2025; To appear in NeurIPS 2025

This paper estabishes a theory of when parallel Newton methods (DEER) efficiently converge.