Kun Wang (王坤)

I am currently a Postdoctoral Fellow at City University of Hong Kong. I received the Ph.D. degree in 2024 from the Department of Computer Science, City University of Hong Kong (CityU), supervised by Prof. Zhenjiang Li. I also work closely with Prof. Zimu Zhou. Before that, I received the B.E. degree in Software Engineering, from Xidian University (XDU), China, 2020.

My research interests primarily lie in designing and building efficient On-Device AI systems (e.g., in Smartphones and NVIDIA Jetson), including:

(1) Optimizing KV-Cache of On-Device LLM systems.
(2) Building efficient On-Device training systems for Heterogeneous Federated Learning (HFL).
(3) Developing efficient On-Device inference engine for Deep Neural Networks (DNNs).

Email  /  Linked In

profile photo
Selected Publications
image 1 [MobiCom' 24] LATTE: Layer Algorithm-aware Training Time Estimation for Heterogeneous Federated Learning

Kun Wang, Zimu Zhou, Zhenjiang Li

[Paper | Slides | Project | Code]

An efficient training time estimation middleware for accelerating FL system convergence on heterogeneous edge devices.
image 1 [TMC' 24] SwapNet: Efficient Swapping for DNN Inference on Edge AI Devices Beyond the Memory Budget

Kun Wang, Jiani Cao, Zimu Zhou, Zhenjiang Li

[Paper | Slides]

An efficient memory management middleware for DNN inference on mobile edge devices, enabling large models deployment beyond the device memory constraints.
image 1 [MobiCom' 23] A Workload-Aware DVFS Robust to Concurrent Tasks for Mobile Devices

Chengdong Lin, Kun Wang, Zhenjiang Li, Yu Pu

[Paper | Slides | Project | Code]

An efficient power governor for scaling processor's frequency in concurrent tasks scenarios on mobile edge devices.
Other Publications
  • [IPSN' 24 Poster] Enhancing Human Motion Sensing with synthesized Millimeter-Waves

    Xiaotong Zhang, Kun Wang, Zhenjiang Li, Jin Zhang

  • [TURC' 23] Poster: Gaze Tracking on Any Surface with Your Phone

    Jiani Cao, Chengdong Lin, Yang Liu, Kun Wang, Zhenjiang Li

  • [TMC' 24] Practical Gaze Tracking on Any Surface with Your Phone

    Jiani Cao, Jiesong Chen, Chengdong Lin, Yang Liu, Kun Wang, Zhenjiang Li
Teaching Assistant
  • Lab Tutor, [CS2310] Computer Programing, 2024 Spring, CityU
  • Lab Tutor, [CS2310] Computer Programing, 2023 Spring, CityU
  • Lab Tutor, [CS2310] Computer Programing, 2022 Spring, CityU
  • Lab Tutor, [CS2204] Fundamentals of Internet Application Development, 2023 Fall, CityU
Professional Services

 Invited Reviewer for:
  • ACM MobiCom, 2025
  • ACM SenSys, 2024-25
  • ACM/IEEE IPSN, 2024
  • ACM/IEEE IWQoS, 2023-24
  • ACM IoTDI, 2022,2024
  • IEEE INFOCOM, 2022-24
  • IEEE ICDCS, 2023-24
  • IEEE Transactions on Mobile Computing (TMC)
  • ACM Transactions on Computer Systems (TOCS)
Selected Honors & Awards
  • Postgraduate Scholarship, CityU, Hong Kong, 2020-2024
  • National Scholarship, Ministry of Education, China, 2019
  • National Scholarship, Ministry of Education, China, 2018
  • The Second Prize Scholarship, Xidian University, China, 2017

Last updated Nov 2024

Thanks to Jon Barron for sharing the code of his personal webpage