Sherjil Ozair
Co-Founder at Project Prometheus
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSherjil Ozair is a machine learning researcher and entrepreneur with 15 years of experience building and deploying generative models and sequence models. Based in San Francisco, he has held research roles at DeepMind, Google, Tesla, and Baidu and now co-leads Project Prometheus while serving as CEO of General Agents. His academic foundation includes a PhD in Machine Learning from Université de Montréal and early work on GANs and unsupervised methods that informed later applied research. Sherjil is also an active open-source contributor—his refactors to the Theano-based Blocks framework and a widely used char‑RNN TensorFlow implementation highlight a blend of deep theoretical insight and pragmatic engineering. Colleagues describe him as someone who both trains cutting‑edge models and restructures core libraries to make them more robust and maintainable. He brings a rare combination of research rigor, production experience, and startup leadership to AI productization.
15 years of coding experience
7 years of employment as a software developer
Indian Institute of Technology Delhi (IIT Delhi)Doctor of Philosophy (PhD) Machine Learning, Doctor of Philosophy (PhD) Machine Learning at Université de Montréal
English, Hindi, French
Stackoverflow
41reputation
769reached
0answers
2questions
Github Skills (16)
neural-network10
machine-learning10
lstm10
rnn-model10
tensorflow10
python10
n10
theano10
natural-language-processing9
refactoring9
nlp9
recurrent-neural-networks9
spatial-query6
r-tree6
indexing6
Programming languages (14)
Github contributions (5)
sherjilozair/char-rnn-tensorflow
Nov 2015 - Jan 2019
Multi-layer Recurrent Neural Networks (LSTM, RNN) for character-level language models in Python using Tensorflow
Role in this project:
ML Engineer
ML EngineerContributions:57 commits, 28 PRs, 50 pushes in 3 years 1 month
Contributions summary:Sherjil primarily contributed to the development of a character-level language model using recurrent neural networks. Their initial commit established the foundational structure. Subsequent commits involved abstracting the model into a separate class, incorporating inference functionality for sampling text, and refactoring the code for better organization. The user also updated and refined the model, including better defaults and the use of the Adam optimizer, showcasing their focus on model design and optimization.
pythonmulti-layertensorflowrnnlstm-neural-networks
mila-iqia/blocks
Dec 2014 - Dec 2014
A Theano framework for building and training neural networks
Role in this project:ML Engineer
ML EngineerContributions:9 commits in 1 day
Contributions summary:Sherjil focused on refactoring and updating core components of the Blocks framework, a Theano-based library for neural network construction. They addressed issues related to initialization and parameter handling within several brick classes. Their contributions included removing redundant code, streamlining initialization configurations, and improving the overall structure of the sequence generator, attention, and recurrent modules.
pytorchdeep-learningtheanoneural-networksmachine-learning
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.