Chengkun Wei

Chengkun Wei

ZJU-100 Talents / QiZhen Young Scholar. AI security, AI infrastructure & systems.

Chengkun Wei

test I am a ZJU-100 Talents Program Researcher and Ph.D. advisor at the College of Computer Science and Technology, Zhejiang University. I am also a Huawei QiZhen Outstanding Young Scholar.

My research focuses on：AI security, AI Infrastructure and Systems，Interdisciplinary AI Applications. I have published many papers as first or corresponding author in top-tier venues such as NDSS, CCS, USENIX Security, IEEE TIFS, IEEE TDSC, SIGCOMM, HPCA, USENIX ATC, and ACL. Representative works include: GradEscape, DC-SGD, Achelous and Triton.

My research is rooted in computer architecture and hardware software co-design. I study the infrastructure and system architecture for modern AI systems, including high-performance networks, high-performance cloud computing systems, AI infrastructure platforms, and Agentic operating systems (Agent OS).

Building upon these system foundations, my work further explores AI security and performance optimization for large-scale AI systems. I am also interested in interdisciplinary AI research, exploring applications of AI in areas such as digital twins, brain-computer systems, and AI-driven social science research.

My research has been supported by the National Natural Science Foundation of China (NSFC), the Ningbo Key R&D Program, the Zhejiang Postdoctoral Fellowship Program, the CCF–Ant Research Fund, and Alibaba Cloud research programs. I received the Zhejiang Provincial Science and Technology Progress First Prize, the NDSS 2024 Distinguished Paper Award, and dual championships in the NeurIPS 2024 LLM Privacy Challenge (attack and defense tracks). My research has also been widely adopted in industry systems and open-source projects

Research Interests

My research focuses on the security, infrastructure, and interdisciplinary applications of large models and intelligent agent systems.

  • AI Security：Access control mechanisms, content safety, agent attack and defense, foundation model security, and privacy-preserving machine learning.
  • AI Infrastructure and Systems：Infrastructure for large-scale AI systems, including AI infrastructure platforms, Agentic operating systems (Agent OS), high-performance networking, and cloud system optimization.
  • Interdisciplinary AI Applications：Applications of AI across disciplines, including digital twins, brain–computer systems, and AI-driven social science research.

Research Highlights

  • Systems & security at the crossroads: Focus on security and infrastructure for large models and intelligent agent systems.
  • Engineering and theory together: Emphasis on both system implementation, performance optimization, and algorithmic foundations.
  • Real-world driven research: Problems are grounded in real systems and applications.
  • Industry–academia collaboration: Partnerships with internet companies, security firms, and key national institutions.

Representative Work

Selected publications appear on the Publications page. For a full list, see Google Scholar.

Honors & Awards

  • Huawei QiZhen Outstanding Young Scholar, 2025
  • Zhejiang Provincial Science and Technology Progress First Prize, 2024
  • NDSS Distinguished Paper Award, 2024 (four papers globally)
  • NeurIPS LLM Privacy Challenge (LLM-PC), Attack Track First Prize, 2024
  • NeurIPS LLM Privacy Challenge (LLM-PC), Defense Track Best Defense Award, 2024
  • National OS Capability Competition, Outstanding Advisor, 2024
  • Ant Academic Star, Ant Group, 2021

Contact