Rupesh Srivastava
Senior Research Scientist at MBZUAI (Mohamed bin Zayed University of Artificial Intelligence)
California, United States
Summary
🤩
Rockstar🎓
Top SchoolRupesh Srivastava is a Senior Research Scientist with 13 years of experience building generative models, reinforcement learning agents, and production-ready ML infrastructure. Based in California, he has led research and software teams at NNAISENSE—where he produced innovations like Upside-Down Reinforcement Learning, Bayesian Flow Networks, and the EvoTorch library—and now contributes to open-source LLMs at MBZUAI’s Institute of Foundation Models in Silicon Valley. His background includes a PhD-level AI training at IDSIA and early industry impact from a Microsoft internship that helped win the 2015 COCO image captioning challenge. A pragmatic engineer, he blends theory and implementation, adding GPU-accelerated neural network support (pycuda/cuDNN) and performant feedforward math to open-source projects. He’s comfortable switching hats between research, engineering leadership, and hands-on backend development, often turning novel ideas into usable code and libraries. Notably, his work marries classical evolutionary methods with deep learning to produce practical generative and RL solutions.
13 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy - PhD Artificial Intelligence, Doctor of Philosophy - PhD Artificial Intelligence at The Swiss AI Lab IDSIA
Github Skills (8)
neural-network10
backend10
cudnn10
python10
pycuda10
numpy10
machine-learning9
hdf4
Programming languages (4)
Github contributions (5)
IDSIA/brainstorm
Mar 2015 - Jan 2017
Fast, flexible and fun neural networks.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:313 commits, 12 PRs, 43 pushes in 1 year 10 months
Contributions summary:Rupesh made significant contributions to the project, primarily by implementing new functionality within the feedforward layer. This involved adding support for new mathematical operations within the numpy handler and implementing these in Python code. Further work added support for pycuda and cuDNN acceleration of neural network layers. Other improvements included modifications to data iterators and general refactoring.
deep-learningneural-networksmachine-learningneural-networktensorflow
nnaisense/evotorch
Aug 2022 - Dec 2024
Advanced evolutionary computation library built directly on top of PyTorch, created at NNAISENSE.
Contributions:38 reviews, 22 PRs, 29 pushes in 2 years 4 months
optimization-algorithmspythonneuroevolutionon-topevolutionary-computation
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.