Trevor Teolis

Postdoctoral Researcher, Rice University
Department of Computational and Applied Mathematics & Operations Research (CMOR)
Email: tt111@rice.edu

Research Overview

My background is in interacting particle systems, collective dynamics, and kinetic and hydrodynamic limits. Around 2023, I became interested in machine learning through transformer token dynamics and particle-based inference (and also because LLMs are cool). Much of my current work is motivated by Bayesian inverse problems: how to model physical systems (the forward model), and how to infer unknown parameters from observations (the inverse problem). On the forward side, I study neural operators for kinetic, fluid, and wave dynamics; on the inverse side, I develop particle-based methods for sampling posterior distributions.

SVGD and alignment simulations · WaveBench acoustic pressure fields

Background

I received my Ph.D. in Mathematics from the University of Illinois Chicago, advised by Roman Shvydkoy. My doctoral work studied collective dynamics and interacting particle systems, including Cucker–Smale, Euler alignment, and Fokker–Planck–Navier–Stokes models, and established rigorous limits from microscopic particle dynamics to kinetic and macroscopic fluid equations.

Upcoming Conferences

About Me

I like to play tennis, soccer, and football (anything with running and a ball is pretty good). Occasionally, I make music with guitar and piano. I tend to inhabit Houston, Milwaukee, Chicago, and D.C. Sometimes, I end up in other places.


Last updated: October 2026