Wenyan Chen

Nanyang Technological University Singapore

prof_pic.jpg

Wenyan Chen (陈文艳)

Email: wenyan.chen at ntu.edu.sg

Greetings! I am a Postdoctoral Researcher in the Hyperscale Systems and Cloud Architecture Lab at Nanyang Technological University (NTU), Singapore, working with Prof. Dmitrii Ustiugov. Previously, I was a PhD student at the Cloud and Distributed Systems (CDS) Lab at University of Macau (2021.08–2025.08), co-supervised by Prof. Huanle Xu (徐欢乐) and Prof. Kejiang Ye (叶可江).

Research Interests

  • GPU virtualization: Improving GPU utilization through temporal and spatial sharing while preserving performance isolation for co-located workloads.
  • LLM inference systems: Accelerating LLM inference with adaptive recomputation and kernel fusion.
  • Multimodal/VLM serving: Balancing accuracy and latency in multimodal inference through token pruning.

news

Jan 30, 2026 eLLM is accepted by EuroSys 2026.
Dec 08, 2025 FedSUV is accepted by INFOCOM 2026.
Sep 27, 2025 FedDance is accepted by SoCC 2025.

selected publications

  1. High Throughput and Low Latency LLM Serving via Adaptive KV Caching
    Wenyan Chen, Chengzhi Lu, Huanle Xu, Kejiang Ye, and Chengzhong Xu
    In proceedings of European Conference on Computer Systems, 2026
  2. Multiplexing Dynamic Deep Learning Workloads with SLO-awareness in GPU Clusters
    Wenyan Chen, Chengzhi Lu, Huanle Xu, Kejiang Ye, and Chengzhong Xu
    In proceedings of European Conference on Computer Systems, 2025
  3. SC
    Interference-aware multiplexing for deep learning in gpu clusters: A middleware approach
    Wenyan Chen, Zizhao Mo, Huanle Xu, Kejiang Ye, and Chengzhong Xu
    In proceedings of the International Conference for High Performance Computing, Networking, Storage, and Analysis, 2023