Kushagra Tiwary
Indian miniature painting style transfer

Download Photo

Kushagra "Kush" Tiwary

Office: 32-G482 & E14-374H
ktiwary [at] mit [dot] edu | LinkedIn | GitHub | X | Google Scholar | CV


I'm a 3rd year PhD Student at MIT co-advised by Boris Katz, Tomaso Poggio, and Ramesh Raskar. I also work closely with Brian Cheung. I received my S.M. in 2023 from MIT, where I worked on 3D vision and imaging, specifically on incorporating realistic physics of light propagation to improve neural rendering. I received my B.S.'19 ECE from University of Illinois at Urbana-Champaign. Prior to MIT, I worked on vision and robotics and built the Software 2.0 stack at a self driving startup called Optimus Ride.

My research builds artificial intelligence that can discover new scientific phenomena and designs. I let AI agents interact and evolve inside simulated worlds grounded by physical laws; intelligence emerges through embodiment and interaction, and with it, discovery. Among the phenomena I study is intelligence itself — what gives rise to its emergence and diversity, and how those principles can inform the design of new artificial intelligence. I am broadly interested in AI, AI for science, simulation and modeling, science of intelligence, and neuroevolution.

I'm also passionate about public engagement and science advocacy. Recently, I wanted to show how AI can be more than just a tool for automation but also used to understand human and animal perception. I've exhibited works at the MIT Museum and the Museum of Science.

I am grateful to be the first in my extended family to be in a PhD program. To learn more about how to apply to PhD programs, please look at the Media Lab's SOS Program and the EECS Graduate Application Assistance Program (GAAP). Reach out to chat or collaborate.

Working Snapshots

Embodied Eyes For Scientific Discovery: Generation and Verification Loops to ask the "why" questions in vision
Redwood Center for Theoretical Neuroscience, UC Berkeley
What-If Machines for Vision: Evolving Eyes and Brains with AI
AI For Science: Imagination in Action @ MIT
Tedx Boston: Can AI Recreate 500 Million Years of Vision Evolution?

NEWS -

PUBLICATIONS +

MY BLOGS +

April 4, 2020: The Perception Problem