Huangwei Wu (吴黄巍)
I am currently a Ph.D. candidate at Shanghai Jiao Tong University, advised by Prof. Meng Jin.
My research focuses on pushing the limits of resource-constrained intelligence by building acoustic-empowered systems for mobile/low-end devices to enable novel applications.
Recently I have also been building on-device LLM agents, and exploring how agents can accelerate acoustic research and the wider scientific workflow.
Feel free to reach out: wuhuangwei@sjtu.edu.cn, whw0315@gmail.com
Education
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University (September 2022 - present)
Phd Student in Information and Signal ProcessingCollege of Electronics and Information Engineering, Sichuan University (September 2018 - June 2022)
B.S. in Electronics and Information Engineering
Publication
* represents first authors, † represents corresponding authors
[CCF-A] Huangwei Wu*, Weiguo Wang†, Meng Jin†, Zhuxuan He, Qi Cao, Xinbing Wang, Chenghu Zhou. “Enabling ‘See-and-Point’ Communication between Robots,” IEEE Transactions on Mobile Computing, 2026.
[CCF-A] Huangwei Wu*, Tingchao Fan*, Meng Jin†, Tao Chen, Xinbing Wang, Chenghu Zhou. “Neural-Enhanced Modulation for Spatial Selective Transmission on Low-End IoT Devices,” IEEE/ACM Transactions on Networking, 2026.
[CCF-A] Tingchao Fan*, Huangwei Wu*, Meng Jin†, Tao Chen, Longfei Shangguan, Xinbing Wang, Chenghu Zhou. “Towards Spatial Selection Transmission for Low-end IoT Devices with SpotSound,” ACM MobiCom, 2023.
Projects
A fully offline call-answering assistant running on a 6 GB entry-level Android phone. Compressing the open-ended call into a finite action space is what makes its 0.8B on-device model reliable at taking calls.
A web-based agent for white papers and literature reviews. Custom workflow where the process is known; ReAct, multi-agent orchestration, and human-in-the-loop where it is not.