Catherine Olsson
Member Of Technical Staff at Anthropic
San Francisco, California, United States
Summary
🤩
RockstarCatherine Olsson is an ML research engineer and AI grantmaker with 13 years of experience bridging computational neuroscience, research software engineering, and programmatic strategy. Currently a Member of Technical Staff at Anthropic, she has a track record of shipping robust tooling and security-minded code across high-profile open-source projects like OpenAI Gym and CleverHans, improving RL environments and adversarial example tooling. Her background at Google Brain, OpenAI, and academic work at NYU and MIT gives her a rare blend of deep research instincts and production-ready engineering skills. She’s moved fluidly between building APIs, hardening ML models, and evaluating AI risks and philanthropy, bringing both technical rigor and policy-oriented perspective to teams. A detail-oriented contributor, she’s often the one fixing subtle compatibility, logging, and test issues that make research code reliable in practice.
13 years of coding experience
7 years of employment as a software developer
English, Swedish, Spanish
Github Skills (32)
json10
environmental10
api-rest10
dev-environment10
adversarial-attacks10
apidoc10
python10
api-design10
restful-api10
machine-learning10
reinforcement-learning10
flask-ask10
enviroment10
security10
api10
Programming languages (9)
Github contributions (5)
openai/gym-http-api
Jun 2016 - May 2017
API to access OpenAI Gym from other languages via HTTP
Role in this project:
Back-end Developer
Back-end DeveloperContributions:69 commits, 33 PRs, 56 pushes in 11 months
Contributions summary:Catherine primarily contributed to the development of the `gym-http-api` repository by implementing core functionalities for the API. Their work included adding features to interact with OpenAI Gym environments, such as listing environments, and getting action/observation space information. These changes involved modifying both the client and server-side code, demonstrating a focus on building out the API's capabilities and ensuring proper data handling and error management. The user also focused on improving the structure, and handling of API keys.
apireinforcement-learningopenailanguagesgym
openai/universe
Dec 2016 - Jan 2017
Universe: a software platform for measuring and training an AI's general intelligence across the world's supply of games, websites and other applications.
Role in this project:Back-end Developer
Back-end DeveloperContributions:10 commits, 10 pushes, 5 comments in 1 month
Contributions summary:Catherine primarily focused on improving the code's maintainability and functionality by adding comprehensive docstrings to the core files. They also allowed URL-encoded tags when using the allocator, improving flexibility. Furthermore, the user removed some deprecated code and fixed URL typos in the README to improve clarity and usability. The user also made changes to logging to improve the information that is displayed.
universegamesmachine-learningtrainingmeasuring
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.