2023–2024
Machine Learning Research Scientist
Leidos
Graph-based knowledge representations for Deep RL Agents.
2025
REM: Evaluating LLM Embodied Spatial Reasoning through Multi-Frame Trajectories
Conference on Language Modeling (COLM) 2025
Jacob Thompson
PhD Student
Language Technologies Institute
Carnegie Mellon University
Bio
I am a PhD student in the LTI at CMU, where I am fortunately advised by Yonatan Bisk. I am interested in augmenting LLMs with cognitive world models so they can:
- learn from a few examples,
- update their knowledge locally without forgetting what they already know, and
- hallucinate less and offer formal guarantees about their reasoning.
Before CMU, I interned at Caltech’s Autonomous Robotics and Control Lab under Soon-Jo Chung, and was first introduced to research at MIT’s Computational Reactor Physics Group.
Publications
Work & research
2019–2020
Computational Reactor Physics Group
MIT
Education
2026–Present
Ph.D., Language Technologies Institute
Carnegie Mellon University
Advisor: Yonatan Bisk
2024–2026
M.S., Robotics
Carnegie Mellon University
2019–2023
A.B., Computer Science + Physics
Harvard University
Contact
first last @ andrew.cmu.edu