Vivek Khandelwal
Member Of Technical Staff - AI Group at The Institution of Engineers (India)
Hyderabad, Telangana, India
Summary
🤩
Rockstar🎓
Top SchoolVivek Khandelwal is an AI-focused software engineer with five years of experience building compiler and ML framework tooling, currently a Member of Technical Staff in AMD’s AI Group in Hyderabad. He has a strong track record contributing to high-profile open-source projects like PyTorch and Torch-MLIR, implementing op lowerings, shape functions, and decompositions that bridge PyTorch with MLIR. Previously an AI compiler engineer at Nod.ai and a short stint at Qualcomm, he combines practical performance engineering with research-grade training from IISc (MTech). Vivek’s work emphasizes end-to-end operator support and benchmarking for real-world models—evident from integrations into SHARK Studio and improvements to ToTMTensor passes—showing a rare mix of low-level compiler expertise and applied model performance tuning.
5 years of coding experience
2 years of employment as a software developer
Master of Technology - MTech, Computer Science, 8.5/10, Master of Technology - MTech, Computer Science, 8.5/10 at Indian Institute of Science (IISc)Bachelor of Engineering - BE, Computer Science, Bachelor of Engineering - BE, Computer Science at Madhav Institute of Technology and Science, Gwalior
English, Hindi
Github Skills (24)
pytorch10
c-language10
python10
operation10
tensorrt10
machine-learning10
mlr10
compiler-compiler10
deep-learning10
tensorflow10
autograd10
compiler10
tensor10
cprogramming-language10
jit9
Programming languages (4)
Github contributions (5)
llvm/torch-mlir
Nov 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:
Back-end Developer
Back-end DeveloperContributions:681 reviews, 120 commits, 565 PRs in 1 year 2 months
Contributions summary:Vivek primarily worked on adding end-to-end support for several PyTorch operations within the MLIR ecosystem. This involved implementing lowering of operations such as `aten.div.Scalar`, `aten.mean`, `aten.numel`, `aten.bitwise_and.tensor`, `aten.zeros`, `aten.index_select`, `torch.arange`, `aten.threshold`, and `aten.index_put.hacked_twin`. The user also decomposed complex ops such as `aten.full` into `aten.empty` and `aten.fill`, and added support for integer inputs for `sum` and `max` operations, indicating a focus on expanding the functionality of the Torch-MLIR project. The contributions include improvements to the ToTMTensor pass.
pytorchmlirtorchcompilerecosystem
nod-ai/SHARK-Studio
May 2022 - Oct 2022
SHARK Studio -- Web UI for SHARK+IREE High Performance Machine Learning Distribution
Role in this project:ML Engineer
ML EngineerContributions:41 reviews, 23 commits, 108 PRs in 5 months
Contributions summary:Vivek implemented features for model benchmarking and integrated new vision models like Squeezenet, Resnet101, and v-diffusion into the SHARK Studio. They added functionalities for running and timing model inference, crucial for performance analysis. The commits demonstrate a focus on integrating and testing different machine learning models within the SHARK framework, enhancing the project's capabilities for high-performance machine learning.
pytorchcudaamdheterogeneousdeep-learning
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.