Research Scientist · GoogleGuest Faculty · IIT Gandhinagar
Making generative models composable, controllable, and reliable —for applications in imaging, vision, and engineering.
Affiliations
Research Scientist
Google — Computational Imaging
Guest Faculty
IIT Gandhinagar
Education
Ph.D., ECE
UIUC
M.S., ECE
Iowa State University
B.Tech., EE
IIT Roorkee
Research Scientist · Google/Guest Faculty · IIT Gandhinagar
Making generative models composable, controllable, and reliable —for applications in imaging, vision, and engineering.
I am a Research Scientist in the Computational Imaging team at Google, where I work with Dr. Peyman Milanfar. I received my PhD in Electrical Engineering from the University of Illinois Urbana–Champaign, advised by Dr. Svetlana Lazebnik. I also hold a guest faculty appointment at IIT Gandhinagar.
My research asks how generative models can become dependable building blocks for recovering, editing, and creating images. At Google, I now focus on making these models reliable for image restoration and editing at scale. My work has shipped in Google products, including Photo Unblur in Pixel and the upscaler in Nano Banana Pro.
Open to Collaboration
I am open to collaborations broadly in the areas of generative models, diffusion models, and image editing. Our team at Google may also have internship openings for PhD students — please reach out via email.
Received the Tech Impact Award at Google (Platforms & Devices) for contributions to the image editing and enhancement stack.
StreetTryOn received the Best Paper Award at the CVFAD Workshop, CVPR 2024. ↗
Recipient of the DAAD AInet Fellowship in Generative Models and Machine Learning. ↗
Awarded the Mavis Future Faculty Fellowship (2023–24) at UIUC (declined).
Recipient of the James M. Henderson Fellowship (2020–21) at ECE, UIUC. ↗
Serving as Area Chair for CVPR 2026 and WACV 2026.
Joined IIT Gandhinagar as Guest Faculty. ↗
UnZipLoRA accepted to ICCV 2025 as a Highlight. ↗
Joined Google as a Research Scientist in the Computational Imaging team.
ZipLoRA accepted to ECCV 2024. ↗
Gemini Team, Google
Technical Report2025
Contributions: Advancing prompt-based image restoration capabilities and building the 4K image upscaler.
Viraj Shah, Ashwini Pokle, Peyman Milanfar, Mauricio Delbracio, Hossein Talebi
In Submission2026
Shubhang Bhatnagar, Ishan Bhatnagar, Viraj Shah, Narendra Ahuja
In Submission2026
Jay Mahajan, Chang Liu, Rauf Makharov, Viraj Shah, Alex Schwing, Svetlana Lazebnik
In Submission2026
Parag Sarvoday Sahu, Ashish Tiwari, Viraj Shah, Shanmuganathan Raman
In Submission2026
Babak Asadi, Peiyang Wu, Mani Golparvar-Fard, Viraj Shah, Ramez Hajj
ECCV2026
Mo Zhou, Keren Ye, Viraj Shah, Kangfu Mei, Mauricio Delbracio, Peyman Milanfar, Vishal M. Patel, Hossein Talebi
TMLR2026
Chang Liu, Viraj Shah, Aiyu Cui, Svetlana Lazebnik
ICCV2025
Viraj Shah, Nataniel Ruiz, Forrester Cole, Erika Lu, Svetlana Lazebnik, Yuanzhen Li, Varun Jampani
ECCV2024
Aiyu Cui, Jay Mahajan, Viraj Shah, Preeti Gomathinayagam, Svetlana Lazebnik
WACV2025
Viraj Shah, Svetlana Lazebnik, Julien Philip
TMLR2024
Viraj Shah, Ayush Sarkar, Sudharsan Krishna, Svetlana Lazebnik
Anand Bhattad, Viraj Shah, Derek Hoiem, David A. Forsyth
Nitish Deshpande, Sanjay Ganapathy, Viraj Shah
Agents4Science Conference2025
Babak Asadi, Viraj Shah, Abhilash Vyas, Mani Golparvar-Fard, Ramez Hajj
Computer-Aided Civil and Infrastructure Engineering2024
Kalina Borkiewicz, Viraj Shah, JP Naiman, Chuanyue Shen, Stuart Levy, Jeff Carpenter
IEEE VIS2021
Selected products I contributed to as a Research Scientist at Google, where research on generative models shipped to users at scale.
Built the 4K image upscaler that ships high-resolution outputs in Nano Banana Pro, and advanced prompt-based image restoration for Gemini 2.5.
[Read more ↗]
Pixel Camera · Google Photos
Developed generative models behind Photo Unblur, sharpening blurry and low-quality shots directly in the Pixel Camera and Google Photos app.
[Read more ↗]Indian Institute of Technology Gandhinagar
A short course covering modern image generative modeling — GANs, diffusion models, and their applications to image synthesis, editing, and inverse problems. The course blends mathematical foundations with hands-on practice on state-of-the-art systems.
WebsiteECE 101 Lab: Intro to Technology for Non-engineers — Instructor
UIUC
ECE 549: Computer Vision — Teaching Assistant
UIUC
ECE 448: Artificial Intelligence — Teaching Assistant
UIUC
CprE 310: Mathematical Foundations of Computer Engineering — Teaching Assistant
Iowa State University
I mentor doctoral students and advise theses on generative models and computer vision — from research direction through publication.
Mo Zhou
Johns Hopkins University · Research Intern at Google
Paper accepted to TMLR.
Babak Asadi
UIUC
Paper accepted to ECCV 2026; 1st prize in the PCI prediction challenge at the Intl. Data Science for Pavements Symposium 2024.
Chang Liu
UIUC
Paper accepted to ICCV 2025.
Nitish Deshpande
UC San Diego
Paper accepted to Agents4Science 2025.
Shrey Sarswat
UIUC · M.S. Thesis
Compositional image generation.
Nidhish Kamath
UIUC · M.S. Thesis
Statistics-aware GANs.
Eric Ji
UIUC · B.S. Thesis
Detecting AI-generated images. Now a Ph.D. student at UIUC.
I have had the privilege of working with and learning from exceptional researchers across academia and industry.
Google (mentor)
Arcade AI (formerly Google)
UIUC (PhD advisor)
UIUC
UIUC
Adobe
Amazon
Amazon
Amazon
Siemens
NYU (MS advisor)