Ruiyang Sun
Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolRuiyang Sun (Ryan) is a senior undergraduate at Peking University double-majoring in Physics and Artificial Intelligence, with five years of hands-on experience in Safe Reinforcement Learning and LLM alignment research under Prof. Yaodong Yang at PKU Pair Lab. His current work investigates emergent socio-dynamic behaviors and misalignment in AI-human mixed ecosystems, focusing on LLMs, large multimodal models, and LLM-powered autonomous agents to promote safer, more harmonious integration of AI into society. As an ML engineer contributor to the widely used OmniSafe SafeRL framework, he refactored core buffer components, added robust tests, and fixed bugs to strengthen infrastructure for safe RL research. He blends technical depth with interdisciplinary thinking—drawing on psychology, sociology, and cognitive science—to study how social learning and cooperation might boost collective intelligence. Outside research he develops Apple apps and is building an AI-powered teamwork research toolkit, signaling a practical bent toward translating research into usable tools.
5 years of coding experience
Senior, Senior at Peking University
Chinese, English
Github Skills (5)
deep-reinforcement-learning10
pytorch10
python10
reinforcement-learning10
testing10
Programming languages (4)
Github contributions (5)
PKU-Alignment/omnisafe
Feb 2023 - Mar 2023
OmniSafe is an infrastructural framework for accelerating SafeRL research.
Role in this project:
ML Engineer
ML EngineerContributions:40 reviews, 9 commits, 18 PRs in 29 days
Contributions summary:Ruiyang primarily focused on refactoring and enhancing the buffer functionalities within the OmniSafe framework, as evidenced by the refactoring of `OnPolicyBuffer` and `VectorOnPolicyBuffer` classes. The changes include implementing various test cases to ensure the robustness of the buffer, and fixing bugs, and enhancing logging. The work involved in depth interaction with the fundamental components for safe reinforcement learning tasks.
pytorchbenchmark-suitedeep-learningreinforcement-learningsafe-reinforcement-learning
rockmagma02/PKUbookpdfDownloader
Jul 2021 - Sep 2022
Contributions:1 release, 9 commits, 2 pushes in 1 year 2 months
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.