Ryan Lowe
Research Ecosystem Lead at Meaning Alignment Institute
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolRyan Lowe is a research leader focused on AI alignment and multi-agent systems, currently leading research ecosystem efforts at the Meaning Alignment Institute after several years on OpenAI’s safety team studying reward modeling. He completed a PhD at McGill/Mila where he worked on emergent communication, multi-agent reinforcement learning, and neural dialogue systems, and has 11 years of experience spanning internships and research roles at Facebook AI, OpenAI, and academic institutes. Ryan pairs deep theoretical work with hands-on engineering—his contributions to the widely used multiagent-particle-envs repo show he improves core environment logic, action/communication handling, and reward callbacks. Based in Berkeley, he brings both leadership in research management and a practiced eye for robust experimental infrastructure, bridging evaluation datasets and production-ready code.
11 years of coding experience
6 years of employment as a software developer
Bachelor of Engineering (B.Eng.) Mathematics and Engineering Robotics Option, Bachelor of Engineering (B.Eng.) Mathematics and Engineering Robotics Option at Queen's UniversityMaster's degree Computer Science, Master's degree Computer Science at McGill University
Github Skills (8)
environmental10
dev-environment10
environ10
multi-agent-reinforcement-learning10
python10
enviroment10
gymnasium9
openai-gym9
Programming languages (3)
Github contributions (5)
openai/multiagent-particle-envs
Aug 2017 - Jul 2018
Code for a multi-agent particle environment used in the paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments"
Role in this project:
Back-end Developer
Back-end DeveloperContributions:35 commits, 3 PRs, 14 pushes in 10 months
Contributions summary:Ryan made several updates to the core environment file, `multiagent/environment.py`, suggesting they were working on the central logic of the multi-agent system. They fixed bugs related to action spaces and communication, indicating a focus on ensuring proper agent interaction and control. Furthermore, they made changes to the interactive policy and scenario files. The user also updated the project dependencies and fixed the reward callback, thus improving the overall project functionality.
actor-criticmulti-agent
ml-retrospectives/neurips2019
Jul 2019 - Dec 2019
Website for the NeurIPS workshop 2019
Contributions:20 commits, 4 PRs, 26 pushes in 4 months
neuripsworkshop
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.