Shreyas Kapur

I am a PhD student at Berkeley AI advised by Prof. Stuart Russell. I am broadly interested in developing algorithms that can learn to construct rich, interpretable models of the world as quickly and robustly as humans do.

I did my undergraduate and Master's at MIT, where I was part of the Computational Cognitive Science group advised by Prof. Josh Tenenbaum. During my time as an undergrad, I interned at DeepMind, Waymo, and worked on research projects under Prof. Pattie Maes at the Fluid Interfaces Group where our work was featured on 60 minutes. I also helped run HackMIT.

Outside computer science, I love writing music, DJing, taking photos, and flying airplanes.


[email]   [twitter]   [github]   [youtube]


Selected Publications

Diffusion On Syntax Trees For Program Synthesis
S. Kapur, E. Jenner, S. Russell
NeurIPS (2024), System 2 Reasoning At Scale Workshop
[Project Page]

Evidence of Learned Look-Ahead in a Chess-Playing Neural Network
E. Jenner, S. Kapur, V. Georgiev, C. Allen, S. Emmons, S. Russell
NeurIPS (2024)
[Project Page]

AlterEgo: A Personalized Wearable Silent Speech Interface
A. Kapur*, S. Kapur*, P. Maes
IUI (2018)
[Video], [Project Page], Press: [1, 2, 3]


Awards

  1. First Place, MIT's Autonomous Robotics Competition MASLAB, 2017.
    Worked on Vision and Planner on the robot.

  2. First Place, MIT's Web Programming Competition web.lab, 2018.
    Worked on a presentation design software that allows the user to make fancy animated presentations right in the browser.

  3. 3rd Grand Award, Intel International Science Fair.
    Built a system to compute eyeglass prescriptions using a mobile app.