Saurav Shekhar
Software Engineer at Google
Zurich, Zurich, Switzerland
Summary
👤
Senior🎓
Top SchoolSaurav Shekhar is a software engineer with 13 years of experience building ML-driven systems and trust-and-safety infrastructure, currently working on YouTube trust and safety at Google in Zurich. He combines strong academic foundations from IIT Kanpur and ETH Zürich with hands-on ML engineering—from handwriting and emoji embedding research to production-scale optimization on AWS SageMaker. His open-source contributions to the widely used ROOT framework include rigorous backpropagation and RNN testing plus CUDA support, reflecting deep knowledge of neural architectures and GPU workflows in scientific contexts. Past internships at CERN, Amazon, and Goldman Sachs demonstrate an ability to move between research and production environments and to deliver robust data-processing and monitoring systems. Known for pragmatic problem-solving, he brings both research rigor and production discipline to complex ML and safety challenges. An unexpected detail: colleagues jokingly describe him as a "highly trained monkey" at Google, hinting at a playful, collaborative personality behind his technical depth.
12 years of coding experience
Indian Institute of Technology KanpurMaster’s Degree, Computer Science, Master’s Degree, Computer Science at ETH Zürich
Hindi, English, German, Sanskrit
Github Skills (9)
cuda10
machine-learning10
rnn-model10
c-language10
deep-learning10
cprogramming-language10
n10
backpropagation10
data-analysis9
Programming languages (5)
Github contributions (5)
root-project/root
Jul 2017 - Feb 2018
The official repository for ROOT: analyzing, storing and visualizing big data, scientifically
Role in this project:
ML Engineer
ML EngineerContributions:17 commits, 6 PRs, 5 comments in 7 months
Contributions summary:Saurav contributed significantly to the backpropagation testing framework within the ROOT project, focusing on the TMVA module. Their work involved adding tests for the DenseLayer, including backpropagation with and without regularization (L1 and L2), as well as tests for the RNN layer. They also added support for Cuda in recurrent propagation. Their contributions demonstrate a strong understanding of neural network architectures and gradient computation, particularly within the context of the ROOT framework's machine learning capabilities.
pythonroot-cernmathematicsc-plus-plusscientific-visualization
sshekh/sshekh.github.io
Feb 2017 - Jul 2021
Github page
Contributions:3 PRs, 57 pushes, 2 branches in 4 years 5 months
js
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.