Parth Savla

San Francisco Bay Area United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Parth Savla is a software engineer with a decade of experience building performant systems and ML-enabled features, currently based in the San Francisco Bay Area. He has driven production improvements at major tech companies through internships at Google and Meta, where his work on personalized promotion UX increased CTR and his Vulkan-backed INT8 inference implementation delivered 30–40% latency wins. Parth is an active contributor to the flagship pytorch/pytorch project, implementing and optimizing Vulkan shaders for quantized tensors and conv/upsample operations to accelerate GPU-backed deep learning. Comfortable across back-end systems, shading languages, and front-end UX, he blends low-level performance optimization with product-minded feature delivery. His academic record (BS CS, 3.9 GPA at UIUC) and prior roles—from AI prototyping to curriculum design—reflect both technical rigor and an aptitude for translating complex problems into usable solutions.

10 years of coding experience

1 year of employment as a software developer

Electrical & Computer Engineering Technologies, Electrical & Computer Engineering Technologies at Middlesex County Vocational Academy Math Science &Engineering Techn

University of Illinois Urbana-Champaign

Spanish, English

Github Skills (14)

tensorrt10

quantization10

vulkan10

pytorch10

tensor10

tensorflow10

glsl10

shader10

gpu10

operation10

cprogramming-language9

c-language9

deep-learning8

machine-learning7

Programming languages (2)

CPython

Github contributions (5)

pytorch/pytorch

Jun 2022 - Jul 2022

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Role in this project:

Back-end Developer

Contributions:1 review, 73 commits, 29 PRs in 1 month

Contributions summary:Parth primarily worked on implementing and improving Vulkan support for PyTorch, focusing on quantized tensors and related operations. They added and optimized Vulkan shaders for quantized operations like `quantize_per_tensor`, `dequantize`, and various arithmetic operations, including `add`, `sub`, `mul`, and `div`. Furthermore, they implemented and benchmarked the `conv2d` and `upsample` operations, specializing in different cases like depthwise and pointwise convolutions to increase the functionality and efficiency of Vulkan-based tensor processing.

pythongpu-accelerationdeep-learninggpunumpy

psavla2/pytorch

Jun 2022 - Jul 2022

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Contributions:37 pushes, 9 branches in 1 month

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.

Request Free Trial