Rohun Agrawal

I am a Computer Science PhD student at Columbia co-advised by Micah Goldblum and Pavel Izmailov. My work is supported by the NSF Graduate Research Fellowship.

Previously, I completed my B.S. in Applied and Computational Mathematics from Caltech where I worked with Georgia Gkioxari on visual reasoning and in Katie Bouman’s lab on imaging inverse problems. I have also interned at Apple training large video models, and at the NASA Jet Propulsion Laboratory calibrating GPR models.

I am interested in developing methods to push the temporal horizons of machine learning models. Specifically, I aim to extend future horizons with world model planning and past horizons with efficient large-scale memory retrieval. I believe both these directions are crucial to achieve truly persistent agents.