Yash Savani
Research Intern at Computer Science Department, Carnegie Mellon
Pittsburgh, Pennsylvania, United States
Summary
👤
Senior🎓
Top SchoolYash Savani is a PhD student at Carnegie Mellon University researching safety, robustness, and efficiency for large generative models under Prof. Zico Kolter, blending theory and systems to make high-dimensional learning methods practical at scale. He connects differential geometry, SDEs, and optimal transport to training and steering techniques—pretraining, fine-tuning, RL, and controlled decoding—while implementing large experiments in PyTorch/JAX, CUDA, Triton and distributed stacks like DeepSpeed and Megatron. With ~11 years of industry and research experience—including roles at Abacus.AI, Primer, and an Adobe research internship—he has shipped AutoML, fairness, GAN/VAE, and forecasting systems and authored multiple papers from production work. His background spans Stanford MS/BS training in statistics and CS and early product engineering (including cofounding a drone startup), giving him a rare mix of rigorous theory, production ML engineering, and systems-level optimization.
11 years of coding experience
5 years of employment as a software developer
IBDP diploma Science, IBDP diploma Science at Dhirubhai Ambani International SchoolDoctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Carnegie Mellon University School of Computer ScienceMaster of Science - MS Statistics, Master of Science - MS Statistics at Stanford University
English, Hindi, Gujarati, French
Github Skills (19)
reproducible-science8
reproducible-research7
cli6
docker6
computer-vision4
lua3
machine-learning2
gluon2
chainer2
caffe2
pytorch2
apollo2
neovim-plugin2
nvim2
neovim2
Programming languages (6)
Github contributions (5)
yashsavani/dotfiles
Oct 2021 - Oct 2023
My Dotfiles
Contributions:1 PR, 77 pushes, 1 branch in 2 years
yashsavani/Advanced-Foundations-of-ML
Oct 2021 - Dec 2021
Contributions:24 commits, 51 pushes, 1 branch in 1 month
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.