Shivangi Singh
Technical Project Manager at CATT Lab (Center for Advanced Transportation Technology Laboratory)
College Park, Maryland, United States
Summary
🤩
Rockstar🎓
Top SchoolShivangi S is an engineer-turned-Product Program Manager with 10 years of experience building and launching B2B and B2C digital products that nudge users toward better decisions while delivering delightful experiences. With a Master's in Software Engineering and a graduate certificate in Project Management from the University of Maryland, she blends strong technical execution with product strategy and cross-functional leadership. She has a track record of delivering complex end-to-end projects, mentoring peers, and aligning engineering efforts to business outcomes. As an active open-source contributor, she improved memory utilization and enabled tensor-parallel MoE features in the widely used DeepSpeed library, showing deep expertise in distributed training and ML performance optimization. Based in College Park, Maryland, she’s passionate about creating technical solutions in education, healthcare, and finance and is open to new opportunities that value product-driven, data-informed engineering.
10 years of coding experience
7 years of employment as a software developer
Bachelor's degree, Computer Science, Bachelor's degree, Computer Science at Motilal Nehru National Institute Of TechnologyKendriya Vidyalaya SangathanMaster's degree, Software Engineering, Master's degree, Software Engineering at University of Maryland
English, Hindi
Github Skills (11)
zero-code10
pytorch10
machine-learning10
deeplearning-ai10
deep-learning10
gpu10
zeroth10
python10
compression9
parallelization8
inference7
Programming languages (6)
Github contributions (5)
deepspeedai/DeepSpeed
Jun 2022 - Aug 2022
DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
Role in this project:
ML Engineer
ML EngineerContributions:12 reviews, 21 commits, 13 PRs in 2 months
Contributions summary:Shivangi primarily contributes to the DeepSpeed library, focusing on enhancements related to Mixture of Experts (MoE) and Zero Redundancy Optimizer (ZeRO) functionalities. They have improved memory utilization, fixed partition ID issues during parameter copying and optimized the interaction between MoE and ZeRO, and enabled contiguous gradients. The user also implemented tensor parallelism for MoE, demonstrating expertise in distributed training.
billion-parametersfinetuningtrainingmixture-of-expertszero
axonn-ai/distrib-dl-tutorial
May 2023 - Nov 2024
Contributions:2 reviews, 21 PRs, 66 pushes in 1 year 6 months
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.