John Agapiou
Research Engineer at google-deepmind
England, United Kingdom
Summary
👤
Senior🎓
Top SchoolJohn Agapiou is a research engineer and data scientist with 11 years of experience blending quantitative neuroscience and computer science to build intelligent systems and practical ML-driven software. With a PhD from UCL and research roles at Rockefeller, Boston University, and Yahoo, he specializes in statistical machine learning and time-series analysis informed by deep domain knowledge of neural data. Now a Staff Research Engineer at Google DeepMind, he contributes to high-profile open-source projects like PySC2 and Melting Pot, improving environment flexibility and multi-agent RL back-ends. He excels at crafting robust, production-ready code for complex simulation environments and enjoys tackling problems that require learning new techniques and creative thinking. A collaborator who values team-driven engineering, he brings both academic rigor and hands-on systems experience to deliver tools that enable real-world utility.
11 years of coding experience
7 years of employment as a software developer
BSc, BSc at University of WarwickUniversity College London
Japanese, German, French
Github Skills (14)
environmental10
dev-environment10
environ10
python10
multi-agent-reinforcement-learning10
reinforcement-learning10
enviroment10
mocker9
mocking9
tensorflow9
mocks9
api8
apim8
api-design8
Programming languages (4)
Github contributions (5)
google-deepmind/meltingpot
Aug 2021 - Jan 2023
A suite of test scenarios for multi-agent reinforcement learning.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:13 releases, 111 reviews, 163 commits in 1 year 5 months
Contributions summary:John primarily contributed to the back-end functionality of the `meltingpot` repository, which focuses on multi-agent reinforcement learning. Their commits demonstrate a focus on improving the bot implementation and scenario logic, ensuring bots function correctly under various conditions. Specifically, they made modifications related to action specifications, bot behavior, and handling of timesteps, indicating a deep understanding of the underlying environment's mechanics. They also addressed code maintenance tasks such as adding lint directives.
multiagent-reinforcement-learningagentreinforcement-learningreinforcement-learning-agentlogistics
google-deepmind/pysc2
Jan 2018 - Apr 2018
StarCraft II Learning Environment
Role in this project:Back-end Developer
Back-end DeveloperContributions:8 commits in 2 months
Contributions summary:John contributed to the PySC2 environment, primarily focused on environment and replay functionality. They removed unused code, added a mock environment for testing, and modified environment specs for multiplayer support. The user also implemented features like sending chat messages and allowing for replay prefix customization, demonstrating a focus on improving environment flexibility and user experience. Additionally, the user fixed a typo and modified port handling, enhancing the robustness of the project.
starcraftstarcraft-iiblizzard-apideep-learningreinforcement-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.