姜伟雄 Weixiong Jiang

同济大学电子与信息工程学院助理教授 Assistant Professor, Tongji University

AI 处理器与软件栈设计

聚焦端侧大语言模型处理器、算子优化、架构探索与软硬件协同设计。

姜伟雄现任同济大学电子与信息工程学院助理教授。博士毕业后先后任职于华为海思图灵架构部与蔚来汽车自动驾驶芯片开发部,长期从事 GPU 与 NPU 的架构定义、方案设计、RTL 开发及性能建模。

AI Processors and Software Stacks

Focused on edge LLM processors, operator optimization, architecture exploration, and software-hardware co-design.

Weixiong Jiang is an Assistant Professor at Tongji University. Before joining academia, he worked in Huawei HiSilicon's Turing Architecture Department and NIO's autonomous driving chip team on GPU/NPU architecture, solution design, RTL development, and performance modeling.

哈尔滨工业大学 学士 2017 中国科学院大学 博士 2022 华为海思、蔚来芯片团队产业经历 端侧 LPU 全栈研究
B.Eng., Harbin Institute of Technology, 2017 Ph.D., UCAS, 2022 Huawei HiSilicon and NIO chip experience Full-stack LPU research

研究方向

  • 端侧大语言模型处理器(LPU)全栈设计
  • 推理框架、算子设计与优化
  • 架构探索与微架构设计
  • 软硬件协同设计

招生说明

欢迎有意报考的硕士、博士同学联系。课题组围绕 LPU 全栈开展有组织科研,强调方法论训练、系统能力培养与产业场景结合。

培养方式覆盖代码能力、系统设计、论文写作与工程落地,适合希望同时提升研究深度与产业能力的同学。

Research

  • End-to-end LPU design for edge large language models
  • Inference frameworks, operator design, and optimization
  • Architecture exploration and microarchitecture design
  • Software-hardware co-design

Recruitment

Prospective M.S. and Ph.D. students are welcome to reach out. The group is organized around a shared LPU full-stack research platform with strong emphasis on methodology, systems thinking, and industry relevance.

The training path covers coding, system design, paper writing, and engineering execution for students aiming to build both research depth and practical capability.

联系

Contact

Email jiangweixiong@tongji.edu.cn Lab Wiki OpenTPU Lab
Xiaohongshu 同济大学做AI芯片的姜伟雄