Rafael Cosman
Founder & CEO at Protogon Research
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolRafael Cosman is a founder and CEO with 13 years of experience building AI-driven products and startups from incubation to market, currently leading Protogon Research in San Francisco where he applies deep-world models to proprietary trading. He combines hands-on machine learning and back-end engineering chops—contributing safety-focused reinforcement learning environments to OpenAI Gym and storage modules for Botkit—to ship reliable systems that bridge research and production. Rafael’s background includes multiple cofounder roles (Archblock, StreetCode Academy) and an early ML internship at Palantir that resulted in a granted patent, reflecting a pattern of technical leadership and impact. He prioritizes building defensible, dataset-driven models and uses trading as a pragmatic testbed to accelerate AI capabilities. Known for pragmatic engineering and thoughtful productization, he also has experience partnering with clinicians on medical imaging at Arterys. Based in the Bay Area with a CS degree from Stanford, he mixes entrepreneurial grit with deep technical fluency.
13 years of coding experience
9 years of employment as a software developer
High School, High School at La Jolla High SchoolBachelor’s Degree Computer Science, Bachelor’s Degree Computer Science at Stanford University
Stackoverflow
2,579reputation
315kreached
9answers
24questions
Github Skills (29)
environmental10
javascript10
dev-environment10
python10
gymnasium10
openai-gym10
storage10
machine-learning10
reinforcement-learning10
mongodb-database10
enviroment10
botkit10
nodejs10
environ10
mongodb10
Programming languages (11)
Github contributions (5)
openai/gym
Jun 2016 - Aug 2016
A toolkit for developing and comparing reinforcement learning algorithms.
Role in this project:
ML Engineer
ML EngineerContributions:25 commits, 31 PRs, 37 comments in 2 months
Contributions summary:Rafael primarily contributed to the development of reinforcement learning environments within the OpenAI Gym repository, specifically focusing on safety-related tasks. Their work included the creation of new environments like those related to interpretability, and environments with decaying rewards and noisy rewards. The user also made modifications to existing environments to enhance their testing capabilities. Additionally, the user worked on improving documentation and descriptions of the environments.
comparingreinforcement-learning-algorithmsdevelopingdeep-learningreinforcement-learning
howdyai/botkit
Dec 2015 - Jan 2016
Botkit is an open source developer tool for building chat bots, apps and custom integrations for major messaging platforms.
Role in this project:Back-end Developer
Back-end DeveloperContributions:61 commits, 31 PRs, 45 pushes in 1 month
Contributions summary:Rafael primarily contributed to the storage functionalities of the Botkit framework. They implemented and documented storage modules using different technologies, including JFS, Firebase, and MongoDB. The user created, modified, and documented storage interfaces, enabling Botkit to persist data related to teams, users, and channels. Furthermore, they improved the consistency of storage interfaces.
alexa-skills-kitbot-applicationbotbotkitfacebook-messenger
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.