Yuyang Du
I am a Postdoctoral Fellow in the Department of Information Engineering at The Chinese University of Hong Kong (CUHK). In 2025, I received my Ph.D. degree in Information Engineering from the same institution, where I was supervised by Prof. Soung Chang Liew. In 2019, I received my bachelor's degree from the School of Electronics Engineering and Computer Science (EECS), Peking University (PKU), under the supervision of Prof. Yiming Lei.
Before my Ph.D., I was an engineer at Huawei 2012 Lab, where I worked on wireless signal processing algorithms for 5G modem power optimization on Kirin 9000, one of the most advanced flagship 5nm SoCs at the time. In 2024-2025, I was a visiting scholar at Harvard University, where I worked on LLM reasoning and efficient inference with Prof. Marinka Zitnik and Prof. Minlan Yu.
I develop intelligent mobile systems, with applications in industrial automation and smart healthcare. Communication and computation form the foundation of my systems. My recent work addresses two key challenges: 1) cost-effective generative AI for resource-constrained edge devices, and 2) reliable, low-latency communication for real-time mobile applications. My toolkit for building such systems includes efficient LLM serving, task-oriented communication, signal processing, network optimization, and robotics.
*Upcoming Itinerary* Starting in Fall 2026, I will be spending six months at the Department of Electrical and Computer Engineering, UC San Diego (UCSD), working with Prof. Nuria Gonzalez-Prelcic. I look forward to fruitful academic discussions and cultural exchanges during my stay in California!
*Openings!* I will join the Department of Electrical and Computer Engineering at The University of Hong Kong (HKU-ECE) as a Research Assistant Professor in 2027 Spring. Our team has multiple openings for postdocs, research assistants, and visiting students in the following directions: 1) AI-powered wireless networking, 2) cost-effective LLM serving at the edge, 3) generative semantic communication, and 4) SDR-driven URLLC systems. Interested candidates are encouraged to drop an email to yuydu@ie.cuhk.edu.hk with CVs. Prospective PhD/MPhil students interested in 2027 Spring or 2027 Fall admission are also welcome to get in touch, and further details will be announced in due course.
News
- 2026-05: Two demo papers (LightGuard, Real-Time CL-SEC) accepted by Mobisys'26.
- 2026-04: Our paper about SFT for networking agents is accepted by IEEE TMLCN.
- 2026-03: CL-SEC, an LLM-powered cross-layer semantic error correction framework, is now on Arxiv.
- 2026-02: KnowGuard is accepted by ICLR'26.
- 2026-01: VeriRAG and SA-OOSC are accepted by ISQED'26 and ICNC'26, respectively.
- 2025-11: Our distributed edge MoE inference framework OD-MoE is available on Arxiv.
- 2025-10: KnowGuard is available on Arxiv; IndusGCC is accepted by NeurIPS'25.
- 2025-09: One paper about Industrial Robotic Network is available on Arxiv.
- 2025-08: LLMind 2.0, SA-OOSC, IndusGCC and VeriRAG are available on Arxiv.
- 2025-07: One paper about Generative Semantic Communication accepted by IEEE TMLCN.
- 2025-06: One paper about MLLM-assisted Data Mining is accepted by ICCV'25.
- 2025-05: LMT++, a MLLM-empowered knowledge distillation framework, is accepted by IEEE TMI.
- 2025-04: Cellular-X, an LLM agent for cellular network, is accepted by Mobisys'25.
- 2025-03: One paper about Industrial IoT Reliability is accepted by IEEE TVT.
- 2025-02: RaC, a SFT framework for networking agents, is accepted by ICNC'25.
- 2025-01: LibraAlign, a DPO framework for ethical alignment, is available on Arxiv.