Salil Desai
Software Engineer at Facebook
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolSalil Desai is a software engineer with nine years of experience specializing in high-performance ML and systems work, currently on the PyTorch Edge Performance team at Meta in New York. He has contributed impactful back-end and ML engineering improvements to the flagship PyTorch repo—adding sparse/quantized linear support, 8-/16-bit index kernels, and Vulkan profiling to boost inference performance on edge devices. His background includes internships at Citadel, SIG, and NASA and he taught algorithms and machine learning courses at Cornell, reflecting both practical and academic depth. Known for bridging low-level performance engineering with ML model needs, he focuses on squeezing latency and memory out of real-world deployments.
9 years of coding experience
1 year of employment as a software developer
Bachelor of Science - BS, Computer Science, 4.213, Bachelor of Science - BS, Computer Science, 4.213 at Cornell UniversityWest Windsor Plainsboro High School South
Github Skills (9)
quantization10
sparse-matrix10
pytorch10
machine-learning10
c-language10
sparse-data10
deep-learning10
cprogramming-language10
vulkan9
Programming languages (5)
Github contributions (5)
pytorch/pytorch
Jul 2021 - Jan 2023
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:69 reviews, 547 commits, 151 PRs in 1 year 5 months
Contributions summary:Salil focused on enhancing the PyTorch library's support for quantized and sparse linear operations within the context of machine learning and deep learning. Their contributions include implementing serialization/deserialization for sparse quantized linear packed parameters, adding a new owned or borrowed vector for QNNPack BCSR indices/values, and adding the ability to include 16bit and 8bit index kernels. They also updated existing test cases and implemented Vulkan profiling capabilities to optimize the performance of the library.
pythongpu-accelerationdeep-learninggpunumpy
salilsdesai/pytorch
Jul 2021 - Sep 2023
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:2 PRs, 396 pushes, 129 branches in 2 years 2 months
pythongpu-accelerationdeep-learninggpuacceleration
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.