Farshid Parizi
Senior Software Engineer at NVIDIA
Seattle, Washington, United States
Summary
👤
Senior🎓
Top SchoolFarshid Parizi is a Senior Software Engineer and Ph.D. candidate from the University of Washington who blends deep research in ubiquitous computing and VR/AR input with production ML systems engineering. He designs novel sensing and tracking solutions at the intersection of hardware and software, with applied experience building interaction techniques for mixed reality at Facebook Reality Labs and in academic research. Transitioning from research to industry, he has delivered MLSys and backend contributions at OctoAI and now NVIDIA, bringing a pragmatic focus on performance and deployability. An active contributor to the TVM compiler project, he implemented Hexagon-targeted schedules and tests for convolutions and improved QNN support—showing hands-on expertise in accelerating deep learning on specialized accelerators. Based in Seattle, he combines four years of industry experience with ongoing Ph.D. work, uniquely positioning him to translate novel input research into scalable, low-level systems. Colleagues describe him as equally comfortable prototyping hardware-aware algorithms and refining compiler-level optimizations for real-world inference.
4 years of coding experience
8 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Electrical and Electronics Engineering, Doctor of Philosophy (Ph.D.) Electrical and Electronics Engineering at University of WashingtonBachelor's degree Electrical and Electronics Engineering, Bachelor's degree Electrical and Electronics Engineering at Sharif University of Technology
Github Skills (9)
compiler10
hexagons10
compiler-compiler10
python10
cprogramming-language9
performance-optimization9
machine-learning9
c-language9
deep-learning9
Programming languages (1)
Github contributions (5)
apache/tvm
Apr 2022 - Dec 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
Backend Developer
Backend DeveloperContributions:19 reviews, 7 commits, 11 PRs in 7 months
Contributions summary:Farshid primarily contributed to the Hexagon backend within the TVM compiler stack, focusing on adding tests and schedules for depthwise convolution and transposed convolution operations. They implemented and refined schedules for the hexagon target, including adding functionality for the conv2d_transpose_nchw. The user also worked on supporting Relax constants in QNN TOPI operations, and updated the AllGather for Disco support. Further, they modified build scripts and test configurations.
metalvulkancompilertensoropencl
farshidsp/tvm
Apr 2022 - May 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Contributions:20 pushes, 5 branches in 1 month
cpugpu-programminggpu-accelerationtvmdeep-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.