Publications

Underlined are students/interns working with me. Co-first authors are marked with *.

Recent Preprints

• Jiayu Xiao, Ye Wang, Zhi Zhang, Jiarui Guo, Kan Liu, Tao Lan, Minchen Yu, “When Shards Fly: Cross-Tensor In-Flight Conversion for Low-Latency, Low-Memory Weight Updates in Disaggregated RL,” in arXiv preprint arXiv:2611.00646.
• Kaiyu Huang*, Yiting Li*, Angyang Gu*, Pinjia He, Minchen Yu, Qingjiang Shi, “AM-Bench: Understanding Agents under Task Underspecification across Models and Harnesses,” in arXiv preprint arXiv:2610.18090.
• Yihao Zheng, Jingzhe Jiang, Dejiang Zhu, Zhiyuan Tan, Yang Tian, Tao Wang, Minchen Yu, “Semantics, Workflows, and Infrastructure: Understanding Agent Serving at Production Scale,” in arXiv preprint arXiv:2609.34432.
• Zhiyuan Tan*, Dejiang Zhu*, Jingzhe Jiang, Yihao Zheng, Yang Tian, Tao Wang, Minchen Yu, “PackServe: SLO-Aware Request Scheduling for Agentic LLM Serving at Scale,” in arXiv preprint arXiv:2609.33224.
• Minchen Yu, “Bridging Agent Semantics with Spot Capacity: An Elastic and Recoverable Service Model,” vision paper in arXiv preprint arXiv:2608.29581.
• Zhexiang Zhang*, Ye Wang*, Yumiao Zhao, Jiayu Xiao, Qianjing Yang, Xiangyu Wang, Jingzhe Jiang, Qizhen Weng, Ruichuan Chen, Shaohuai Shi, Yin Chen, Minchen Yu, “Janus: Disaggregating Attention and Experts for Scalable MoE Inference,” in arXiv preprint arXiv:2512.13525.

Refereed Papers

• Zhexiang Zhang, Minchen Yu, Yifan Sun, Xu Bai, Xingliang Yuan, Adel N. Toosi, “PixelFlow: Token-Level Workload Management for SLO-Aware and Resource-Efficient Mixed-Resolution DiT Serving,” in the 33rd IEEE International Symposium on High-Performance Computer Architecture (HPCA’27), Salt Lake City, UT, March 2027.

• Yihao Zheng, Zhiyuan Tan, Yicheng Huo, Bohui Wu, Yiyun Zheng, Minchen Yu, “DuetScale: Scaling Disaggregated LLM Serving with Joint Parallelism and Startup Optimization,” in the Symposium on Rigorous LLM Systems (RigoLLM’26, ATC’26 Workshop), Hong Kong, November 2026.
• Yaohua Xi, Minchen Yu, “SemServe: A Systematic View of Exploiting Application Semantics in Agentic Serving,” in the Symposium on Rigorous LLM Systems (RigoLLM’26, ATC’26 Workshop), Hong Kong, November 2026.
• Yiyun Zheng, Lingyun Yang, Yinghao Yu, Guodong Yang, Liping Zhang, Minchen Yu, “Model-as-a-Service in the Wild: Characterizing Production LLM Inference Clusters at Scale,” in 2026 ACM SIGOPS Annual Technical Conference (ATC’26), Hong Kong, November 2026.
• Kaiyu Huang, Xingyu Wang, Mingze Kong, Zhubo Shi, Yuqian Hou, Hong Xu, Zhongxiang Dai, Minchen Yu, Qingjiang Shi, “UniScale: Adaptive Unified Inference Scaling via Online Joint Optimization of Model Routing and Test-Time Scaling,” in the 43rd International Conference on Machine Learning (ICML’26), Seoul, South Korea, July 2026.
• Wenhao Zou, Zhijie Cai, Minchen Yu, Zongshuai Zhang, Guangxu Zhu, “Think in Cloud, Look at Edges: Semantic-Driven Query Decomposition for Efficient Video Reasoning,” in the 43rd International Conference on Machine Learning (ICML’26 Spotlight), Seoul, South Korea, July 2026.
• Kaiwen Chen, Xin Tan, Minchen Yu, Jingzong Li, Hong Xu, “ReasonCache: Accelerating Large Reasoning Model Serving through KV Cache Sharing,” in IEEE/ACM International Symposium on Quality of Service (IWQoS’26), Istanbul, Turkey, June 2026.
• Minchen Yu, Ao Wang, Bohui Wu, Yuxuan Liu, Dong Chen, Haoxuan Yu, Wei Wang, Ruichuan Chen, Dapeng Nie, Haoran Yang, Yu Ding, “Enabling Low-Latency, GPU-Efficient Serverless Inference with Model Swapping,” in ACM Transactions on Architecture and Code Optimization (TACO), 2026.
• Minchen Yu*, Rui Yang*, Chaobo Jia, Zhaoyuan Su, Sheng Yao, Tingfeng Lan, Yuchen Yang, Yue Cheng, Wei Wang, Ao Wang, Ruichuan Chen, “FaaScale: Unlocking Fast LLM Scaling for Serverless Inference,” in the 9th Annual Conference on Machine Learning and Systems (MLSys’26), Bellevue, WA, May 2026.
• Hao Wu, Yaochen Liu, Minchen Yu, Qizhen Weng, Junxiao Deng, Yue Yu, Hao Fan, Song Wu, Wei Wang, and Hai Jin, “Efficient Data Passing for Serverless Inference Workflows: A GPU-Centric Approach,” in ACM European Conference on Computer Systems (EuroSys’26), Edinburgh, UK, April 2026.

• Minchen Yu, Wei Wang, Yue Cheng, Hong Xu, “Cloud-Native Systems for Generative AI Applications: Current Trends and Open Challenges,” Grand Challenges paper in the 1st Frontier AI Systems Workshop (FAISys’25), Hong Kong, November 2025.
• Kaiyu Huang, Hao Wu, Zhubo Shi, Han Zou, Minchen Yu, Qingjiang Shi, “AdaSpec: Adaptive Speculative Decoding for Fast, SLO-Aware Large Language Model Serving,” in the Proceedings of ACM Symposium on Cloud Computing (SoCC’25), Virtual Conference, November 2025.
• Suyi Li, Hanfeng Lu, Tianyuan Wu, Minchen Yu, Qizhen Weng, Xusheng Chen, Yizhou Shan, Binhang Yuan, Wei Wang, “Toppings: CPU-Assisted, Rank-Aware Adapter Serving for LLM Inference,” in USENIX Annual Technical Conference (ATC’25), Boston, MA, July 2025.
• Minchen Yu, Ao Wang, Dong Chen, Haoxuan Yu, Xiaonan Luo, Zhuohao Li, Wei Wang, Ruichuan Chen, Dapeng Nie, Haoran Yang, Yu Ding, “Torpor: GPU-Enabled Serverless Computing for Low-Latency, Resource-Efficient Inference,” in USENIX Annual Technical Conference (ATC’25), Boston, MA, July 2025.[code]

• Minchen Yu, Tingjia Cao, Wei Wang, Ruichuan Chen, “Pheromone: Restructuring Serverless Computing with Data-Centric Function Orchestration,” in IEEE/ACM Transactions on Networking (TON), 2024.
• Minchen Yu, Tingjia Cao, Wei Wang, Ruichuan Chen, “Following the Data, Not the Function: Rethinking Function Orchestration in Serverless Computing,” in the Proceedings of the 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI’23), Boston, MA, April 2023. [code]
• Minchen Yu, Zhifeng Jiang, Hok Chun Ng, Wei Wang, Ruichuan Chen, and Bo Li, ‘‘Gillis: Serving Large Neural Networks in Serverless Functions with Automatic Model Partitioning,’’ in the Proceedings of the 41st IEEE International Conference on Distributed Computing Systems (ICDCS’21), Virtual Conference, July 2021. (Best Paper Runner Up) [code]
• Huangshi Tian, Minchen Yu, and Wei Wang, ‘‘CrystalPerf: Learning to Characterize the Performance of Dataflow Computation through Code Analysis,’’ in the Proceedings of USENIX Annual Technical Conference (ATC ’21), Virtual Conference, July 2021.
• Minchen Yu, Yinghao Yu, Yunchuan Zheng, Baichen Yang, and Wei Wang, “RepBun: Load-Balanced, Shuffle-Free Cluster Caching for Structured Data,” in the Proceedings of IEEE INFOCOM’20, Virtual Conference, July 2020.
• Chengliang Zhang, Minchen Yu, Wei Wang, and Feng Yan, “Enabling Cost-Effective, SLO-Aware Machine Learning Inference Serving on Public Cloud,” in IEEE Transactions on Cloud Computing (TCC), 2020.
• Chengliang Zhang, Minchen Yu, Wei Wang, and Feng Yan, “MArk: Exploiting Cloud Services for Cost-Effective, SLO-Aware Machine Learning Inference Serving,” in the Proceedings of USENIX Annual Technical Conference (ATC’19), Renton, WA, July 2019.
• Huangshi Tian, Minchen Yu, and Wei Wang, “Continuum: A Platform for Cost-Aware, Low-Latency Continual Learning,” in the Proceedings of ACM Symposium on Cloud Computing (SoCC’18), Carlsbad, CA, October 2018.