Sign In

Gaurav Menghani

Staff Software Engineer at Google Research (Deep Learning, On-Device Machine Learning, Computer Vision)

Gaurav Menghani is a Staff Software Engineer at Google Research, where he leads several research projects focused on optimizing large ML models for efficient training and inference across various devices, from microcontrollers to Tensor Processing Unit (TPU)-based servers.12 His work has impacted over 1 billion active Google users across platforms like YouTube, Cloud, Ads, and Chrome.1

Menghani has over 11 years of experience in the tech industry.2 Before joining Google in 2017, he worked at Facebook for 4.5 years, contributing to Facebook's Search system and large-scale distributed databases.13 His research interests include Machine Learning, Algorithms, and Distributed Systems, with a current focus on Efficient Deep Learning.4

Academically, Menghani holds an M.S. in Computer Science from Stony Brook University (State University of New York).13 He is also an author of an upcoming book with Manning Publication on Efficient Deep Learning.1

Menghani is known for his contributions to the field of efficient deep learning. He has published a highly cited survey paper titled "Efficient deep learning: A survey on making deep learning models smaller, faster, and better" in ACM Computing Surveys.5 He has also been involved in various research projects and has several patents to his name, including one for "Application development platform and software development kits that provide comprehensive machine learning services".5

Highlights

Aug 24 · linkedin

I often joke that part of my compensation is the access to the incredibly reliable hordes of TPUs that me and my colleagues have at our disposal at Google Research, but that joke is grounded in reality. Having built a workstation-level deep learning machine on my own in the past year, I know how non-trivial it is to even use off-the-shelf components for a semi-serious use case. Let alone designing a custom accelerator from the ground up for inference and training, developing compilers, training frameworks, networking technologies etc. from scratch and so on and scaling all of that to a gargantuan level.

All this work remains behind the curtains and has been under-appreciated for a while, so I am absolutely chuffed that someone decided to peek into all this crazy impressive work done by so many stellar people who didn't shy away from a challenge.

Related Questions

What are some of the key projects Gaurav Menghani has led at Google?
How has Gaurav Menghani's work impacted Google users?
What is the upcoming book by Gaurav Menghani about?
What was Gaurav Menghani's role at Facebook?
What are Gaurav Menghani's main research interests in machine learning?
Gaurav Menghani
Gaurav Menghani, photo 1
Gaurav Menghani, photo 2
Add to my network

Location

Mountain View, California, United States