Christopher Olah
Cofounder, Interpretability Research Lead at Anthropic
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolChristopher Olah is a research-driven engineering leader with 17 years of experience building and explaining deep learning systems, currently cofounder and Interpretability Research Lead at Anthropic. He led seminal interpretability work at OpenAI and Google Brain — including the circuits project and discovery of multimodal neurons — and combines hands-on engineering with clear, visual explanations of complex models. A prolific open-source contributor, his code and writing (e.g., lucid and a long-standing neural nets blog) advance tools and intuition for model visualization and analysis. Based in San Francisco, he blends rigorous research, practical engineering, and a knack for communicating subtle concepts, often surfacing unexpected structure in neural networks that guide safer, more interpretable AI.
17 years of coding experience
6 years of employment as a software developer
University of Real Life Experience
French, Latin, English
Stackoverflow
416reputation
11kreached
3answers
0questions
Github Skills (30)
python10
architecture10
command-line-interface10
topology10
machine-learning10
3d-printing10
3d-models10
tensorflow10
command-line10
neural-network10
architectures10
haskell10
data-structure9
debug9
algorithm9
Programming languages (8)
Github contributions (5)
tensorflow/lucid
Jan 2018 - Jan 2021
A collection of infrastructure and tools for research in neural network interpretability.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1 release, 3 reviews, 199 commits in 3 years
Contributions summary:Christopher primarily contributed to the codebase by fixing issues related to string formatting and objective functions within the `optvis` directory. They implemented the addition of `misc.show` for displaying images. They also introduced the `ChannelReducer` class, a helper for dimensionality reduction of tensors, and restructured `optvis.param` into multiple submodules, indicating an involvement in the core functionality and architecture of the library.
pytorchinterpretabilitydeep-learningjupyter-notebookinfrastructure
colah/colah.github.io
Mar 2014 - Sep 2022
Role in this project:Data Scientist
Data ScientistContributions:246 commits, 27 PRs, 161 pushes in 8 years 7 months
Contributions summary:Christopher primarily updated a blog post related to neural networks, manifolds, and topology. The contributions include updating the content of the article, adding new figures, clarifying text, and publishing the post to the site. The user is focused on explaining concepts related to neural networks.
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.