Gaurav Shukla
Senior Software Development Engineer at AMD
Karnataka, India
Summary
🤩
Rockstar🎓
Top SchoolGaurav Shukla is a machine learning compiler engineer and principal-level developer with nine years of experience bridging compiler infrastructure and ML systems, currently based in Bengaluru. He has driven production-focused compiler work at GlobalFoundries and AMD and contributed substantial PyTorch-to-MLIR lowering in the high-profile llvm/torch-mlir project, implementing end-to-end support for many aten ops. Gaurav’s background spans full-stack ML web tooling—adding performant Stable Diffusion and ResNet/Albert visualizations to SHARK-Studio—and low-level compiler contributions to LLVM and TensorFlow MLIR. He holds an M.Tech from IIT Delhi and a track record of moving research-grade compiler features into usable toolchains. Known for pragmatic refactors and folding optimizations, he combines deep systems insight with product-minded engineering to accelerate ML deployment.
9 years of coding experience
5 years of employment as a software developer
Indian Institute of Technology Delhi (IIT Delhi)High School, PCM, High School, PCM at Jawahar Navodaya Vidyalaya - JNVBachelor’s Degree, Computer Science & Engineering, Bachelor’s Degree, Computer Science & Engineering at University Institute of Technology, RGPV
English, Hindi
Github Skills (12)
webui10
compiler10
pytorch10
machine-learning10
gradio10
compiler-compiler10
stable-diffusion10
lg10
python10
mlr10
deep-learning9
numpy7
Programming languages (5)
Github contributions (5)
nod-ai/SHARK-Studio
Sep 2022 - Jan 2023
SHARK Studio -- Web UI for SHARK+IREE High Performance Machine Learning Distribution
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:112 reviews, 81 commits, 235 PRs in 4 months
Contributions summary:Gaurav introduced a web interface for the SHARK models, utilizing the Gradio platform. They implemented the web visualization for Resnet50 and Albert_Maskfill models. The user added a V-Diffusion model web visualization to the shark web. Furthermore, the user enhanced the web UI by integrating a stable-diffusion model, adding standard output display, and incorporating instructions for launching the shark-web. The user also optimized the VAE model which improved performance significantly.
pytorchcudaamdheterogeneousdeep-learning
llvm/torch-mlir
Oct 2021 - Jan 2023
The Torch-MLIR project aims to provide first class support from the PyTorch ecosystem to the MLIR ecosystem.
Role in this project:ML Engineer
ML EngineerContributions:85 reviews, 34 commits, 95 PRs in 1 year 2 months
Contributions summary:Gaurav implemented and integrated support for several PyTorch operations within the MLIR framework for the Torch-MLIR project. They added end-to-end (E2E) support for operations such as `aten.view`, `aten.expand`, `aten.mul.Scalar`, `aten.addmm`, `aten.squeeze`, `aten.gt.Scalar`, `aten.where.self`, `aten.ceil`, `aten.zeros`, `aten.ones`, `aten.empty`, `aten.empty_like`, `aten.squeeze.dim`, `aten.eq.Tensor`, `aten.lt.Tensor`, `aten.eq.Scalar`, `aten.lt.Scalar` and `aten.Hardsigmoid` operations. The contributions involved lowering PyTorch operations into the Linalg dialect and, in some cases, refactoring the code for binary comparison ops. They also worked on folding trivial cases for `aten.to.dtype` and `aten.view`.
pytorchmlirtorchcompilerecosystem
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.