Skip to main content

Branton DeMoss


Branton joined the ORI in fall 2021 as a DPhil student with joint affiliations to the GOALS and A2I labs. His academic interests lie in the intersection of deep-learning based world modeling with classical planning, to build agents that can think ahead to act in the world. He is especially interested in how ideas from counterfactual analysis and game theory can be used to make learning-based multi-agent systems more effective.

Before joining the ORI, Branton founded a startup to help researchers explore the literature more effectively, and before that he worked as a machine learning research engineer at different startups working on geophysical modeling and self-driving cars. Before that, Branton studied physics and mathematics at the University of Colorado in Boulder, his home town, where he worked on the DUNE experiment, amongst other research activities.

When not browsing the arXiv, he enjoys rock climbing, riding his bike, and playing the game of go. He also recently restarted the University Go Society, and would love to see you at a meeting and play a game together! You can find his research website here, and his personal site here.