Arun Mallya
Research Scientist at Meta
San Jose, California, United States
Summary
🤩
Rockstar🎓
Top SchoolArun Mallya is a research scientist with 14 years of experience specializing in computer vision and generative AI, currently contributing to Meta's MovieGen video-generation research after a multi-year tenure at NVIDIA where he helped found the Deep Imagination research group and worked on visual generative systems for NVIDIA Edify. He holds a PhD and MS from UIUC with perfect GPAs and a B.Tech from IIT Kharagpur, reflecting strong theoretical foundations paired with production-focused research. Arun is an active open-source contributor to core ML tooling—he's made stability and performance fixes to PyTorch's nn module and enhanced demos and analytics for NVlabs' FUNIT image-translation project—demonstrating comfort across low-level library work and user-facing demos. Based in San Jose, he blends rigorous academic training with practical engineering, often tackling numerical stability and UX details that are easy to overlook but crucial in deployed AI systems.
14 years of coding experience
6 years of employment as a software developer
University of Illinois Urbana-ChampaignBachelor of Technology (B.Tech.), Computer Science and Engineering, 9.49/10.00, Bachelor of Technology (B.Tech.), Computer Science and Engineering, 9.49/10.00 at Indian Institute of Technology, Kharagpur
Stackoverflow
3,441reputation
1.0mreached
1answer
8questions
Github Skills (24)
pytorch10
machine-learning10
lua10
html10
neural-network10
google-tag-manager10
javascript9
performance-optimization9
optmization9
optimisation9
optimization9
website-design8
site-design8
algorithms8
user-interface7
Programming languages (7)
Github contributions (5)
NVlabs/FUNIT
May 2019 - Jul 2019
Translate images to unseen domains in the test time with few example images.
Role in this project:
Front-end Developer
Front-end DeveloperContributions:7 commits, 10 pushes in 2 months
Contributions summary:Arun primarily updated HTML files within the `docs` directory, indicating a focus on the project's documentation and demo interfaces. Their contributions involve incorporating Google Tag Manager for analytics tracking and modifying HTML elements. These changes include adding links, updating the appearance of the demo page and managing external resources like JavaScript files, thus enhancing user experience and data collection.
unseenimage-synthesiscomputer-visiontranslatedomains
torch/nn
Sep 2016 - Oct 2016
Role in this project:ML Engineer
ML EngineerContributions:6 commits, 2 PRs, 6 comments in 27 days
Contributions summary:Arun made several contributions to the `nn` module within the PyTorch library. Primarily, the user modified and enhanced the `MultiLabelSoftMarginCriterion` class, addressing stability issues in the loss and gradient calculations and adding the ability to apply weights to the loss function. Additionally, the user optimized code in `SpatialUpSamplingBilinear.lua` by switching from `reshape` to `resize` and specified output height and width.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.