Oswin So θιζ (θι½ζ)
Safety for Autonomous Systems using Control + Machine Learning @ REALM, MIT AeroAstro

Iβm Oswin So, a 3rd year grad student in REALM at MIT, advised by Chuchu Fan. Previously, I did my undergrad at Georgia Tech, where I was very fortunate to do undergraduate research with Evangelos Theodorou and Molei Tao.
Iβve previously interned at Toyota Research Institute, where I worked on game theoretic planning. I also worked at Aurora as a Behavior Planning Intern during the summer of 2021 under Paul Vernaza and Arun Venkatraman, working on cost function learning via on-policy negative examples for autonomous driving.
See my full CV here (updated February 2025).
Contact: oswinso [at] mit [dot] edu
Follow:
Google Scholar
|
LinkedIn
|
oswinso
|
@oswinso
News
Apr 2025 |
Happy to announce that my papers on constructing safe distributed controllers by combining epigraph form with multi-agent reinforcement learning and improving constraint satisfaction of sampling-based MPC (such as MPPI / CEM) with neural CBFs and better sampling techniques has been accepted to RSS 2025! Stay tuned for the project page for both papers. |
---|---|
Jan 2025 |
DGPPO, our work on learning and using discrete graph CBFs for safe MARL has just been accepted to ICLR! Check out the project page! |
Dec 2024 |
GCBF+, our work on constructing control barrier functions that proves theoretical guarantees on safety for any swarms with any number of agents has just been accepted to T-RO! Check out the project page! |
Oct 2024 |
Happy to announce that our paper solving optimal control problems with reach-avoid constraints using Deep RL has been accepted to NeurIPS 2024! Check out our project page! |
Jul 2023 |
My first paper after joining MIT as a grad student on combining Deep RL and optimal control to synthesize safe, stabilizing controllers has been accepted to RSS 2023! Check out the project page for cool visualizations and videos. |