About

Biography

I am an AI Research Lead working on self-evolving agent infrastructure and agent societies. My research focuses on reinforcement learning, multi-agent systems, and large language model agents, with the goal of building trustworthy, scalable agent systems that can learn, coordinate, and act in complex real-world environments.

I hold a Ph.D. in Engineering from the University of Toronto (2025), an M.Eng. from McGill University, and a B.Eng. in Vehicle Engineering from Beijing Institute of Technology, where I graduated first in my major. During my academic training, I also conducted research internships at UC Berkeley and Mila – Quebec AI Institute.

My work spans long-horizon memory and experience management, adaptive orchestration, multi-agent coordination, execution-grounded evaluation, and real-world deployment. I have authored or co-authored 40+ papers across journals, conferences, and workshops, including ACL, EMNLP, WWW, ICDE, ACM Multimedia, ICRA, IEEE IV/ITSC/ICME, WACV, and NeurIPS/ICML/ICLR workshops, as well as journals such as INFORMS Journal on Computing, IEEE Network, Expert Systems with Applications, and IET Intelligent Transport Systems. My research has received 2,000+ Google Scholar citations with an h-index of 25+. I have also supervised student research published at IEEE IROS and IEEE Transactions on Image Processing, among others.

I maintain active collaborations with researchers at the University of Toronto, McGill University, Tsinghua University, Peking University, MIT, the University of Wisconsin–Madison, Google DeepMind, Meta, and other academic and industry groups. Beyond academia, I have contributed to research translation and AI product innovation through collaborations and advisory work with companies and AI ventures including Momenta, Megvii, Sinovation Ventures, QCraft, XtalPi, BioMap, Skywork AI, Mlion.ai, Lessie.ai, Effyic.com, Autoagents.ai, Gradient HQ, and other AI-related organizations. These experiences shape my broader goal of connecting academic research, open benchmarks and infrastructure, industry partnerships, and real-world evaluation to build deployable agentic AI systems.

Currently, my work centers on collaborative intelligence and the infrastructure required for self-evolving AI systems. I am particularly interested in how agents can accumulate experience, specialize into roles, coordinate with other agents, improve through verified feedback, and operate under human oversight. My long-term vision is to build a transparent, scalable, and participatory ecosystem for increasingly general-purpose and accountable agentic intelligence.

Research Interests

My research interests span several areas at the intersection of artificial intelligence, agent infrastructure, and collective intelligence:

  • Self-Evolving Agent Infrastructure: Building systems that support long-horizon experience management, memory consolidation, adaptive inference, workflow optimization, and recursive improvement.
  • Multi-Agent Systems and Agent Societies: Studying coordination, cooperation, competition, delegation, and role specialization among heterogeneous agents in shared environments.
  • Reinforcement Learning and Autonomous Driving: Designing algorithms for safe, efficient, and comfortable decision-making in mixed traffic and connected autonomous vehicle systems.
  • Evaluation, Benchmarks, and Trustworthy AI: Developing execution-grounded benchmarks, safety constraints, human-in-the-loop evaluation protocols, and accountability mechanisms for deployed agentic systems.
  • Domain-Facing Agent Applications: Translating agentic AI into intelligent mobility, financial analysis, scientific discovery, digital-expert workflows, generative modeling, and industrial operations.
  • Research-to-Deployment Ecosystems: Connecting academic research, open-source communities, startups, and industrial partners to turn real-world constraints into reusable research infrastructure.

Professional Service

Reviewer

  • NeurIPS, ICML, ICLR
  • CVPR, ICCV, EMNLP, AAAI, ICRA, ACL, NAACL
  • IEEE Transactions on Intelligent Transportation Systems
  • IEEE Transactions on Vehicular Technology
  • Transportation Research Part C: Emerging Technologies
  • Scientific Reports
  • Nature Communications
  • TRB Annual Meeting – Transportation Research Board
  • IEEE International Conference on Intelligent Transportation Systems
  • IEEE Intelligent Vehicles Symposium
  • ACM SIGKDD Conference on Knowledge Discovery and Data Mining
  • NeurIPS Workshop on Machine Learning for Autonomous Driving

Skills

  • Programming Languages: Python, C/C++, MATLAB
  • Machine Learning Frameworks: PyTorch, TensorFlow
  • Tools: LaTeX, Git, Docker, ROS
  • Languages: English (Fluent), Mandarin (Native)