Tejan Karmali

I am a PhD student in Computer Science at Stanford, advised by Prof. Kayvon Fatahalian. My work sits at the intersection of reinforcement learning, generative models, and computer graphics. I am most interested in world models: specifically, how do we learn representations that let an agent simulate, plan, and act quickly inside rich 3D environments? Right now I am pursuing this question in the context of fast solutions to 3D games.

Before Stanford, I was a Research Engineer at Avataar working on image-to-3D reconstruction, and a Pre-Doctoral Researcher at Google Research India, where I worked on super-resolved segmentation of satellite imagery with Dr. Varun Gulshan and Alex Wilson.

I did my M.Tech. (Research) at the Indian Institute of Science, in the Video Analytics Lab, advised by Prof. R. Venkatesh Babu and Dr. Varun Jampani. My thesis was on self-supervised representations for landmark estimation and generative image synthesis.

Recent News

[Jan 26]Teaching Assistant for CS248A — Computer Graphics.
[Sep 24]Started PhD in CS at Stanford University.
[Aug 24]Our work AttributeDiffusion, which generates diverse-attributed images, accepted at WACV 2025!
[Jul 24]Our work PreciseControl, which combines the best of StyleGAN and Text-to-Image Diffusion Models for precise attribute control, accepted at ECCV 2024!
[Nov 23]Serving as a reviewer in CVPR 2024.
[Oct 23]Started as a Research Engineer at Avataar!
[Feb 23]Our work NoisyTwins, on long-tail and few-shot image generation for StyleGANs, accepted at CVPR 2023!

[Sep 23]Serving as a reviewer in ICLR 2024.
[May 23]Serving as a reviewer in NeurIPS 2023.
[Nov 22]Serving as a reviewer in CVPR 2023.
[Oct 22]Attended ECCV'22 at Tel Aviv to present HSR and gSR.
[Oct 22]Defended my M.Tech. (Research) thesis!
[Jul 22]Our papers on GAN regularizers (HSR and gSR) accepted at ECCV 2022!
[Jun 22]Our work FLAME, enabling few-shot image editing using StyleGAN, accepted at ACM MM 2022!
[Jun 22]Recognized as Outstanding Reviewer at CVPR 2022.
[Apr 22]Started as Pre-Doctoral Researcher at Google Research India under the guidance of Dr. Varun Gulshan.
[Mar 22]Gave colloquium and submitted thesis on Landmark Estimation and Image Synthesis Guidance using Self-Supervised Networks at IISc.
[Jan 22]Virtually attended WACV 2022 to present LEAD.
[Oct 21]Our work LEAD, a self-supervised technique for landmark estimation, accepted at WACV 2022!
[Jul 21]Our work CSPNet, proposing implicit representation for open and closed 3D shapes, accepted at ICCV 2021!