Xinchen Wan
Network Engineer and Researcher, ByteDance, Seattle, WA

Office: 601 108th Ave NE
Bellevue WA 98004
I am a Network Engineer and Researcher at ByteDance, working on building next-generation AI network. I received my Ph.D. degree in Computer Science and Engineering from Hong Kong University of Science and Technology (HKUST) in 2025, supervised by Prof. Kai Chen. Before that, I received my B.Eng degree in Computer Science from Huazhong University of Science and Technology (HUST) in 2018, supervised by Prof. Song Wu.
Research Interests
- AI Network
- Machine Learning System
- Hardware Acceleration
- Datacenter Networking
News
Jul 12, 2025 | Our short paper Hermod is accepted at SIGCOMM 2025. |
---|---|
Jul 12, 2025 | One co-authored paper MixNet is accepted at SIGCOMM 2025. |
Apr 25, 2025 | One co-authored paper Pallas is accepted at ATC 2025. |
Feb 11, 2025 | PhD thesis defense passed! |
Jan 28, 2025 | One co-authored paper Harmonia is accepted at ASPLOS 2025. |
Selected Publications
- __SIGCOMM__MixNet: A Runtime Reconfigurable Optical-Electrical Fabric for Distributed Mixture-of-Experts TrainingIn Proceedings of the ACM Special Interest Group on Data Communication (SIGCOMM) , 2025
- ____ATC____Towards Optimal Rack-scale μs-level CPU Scheduling through In-Network Workload ShapingIn Proceedings of USENIX Annual Technical Conference (ATC) , 2025
- __ASPLOS__Harmonia: A Unified Framework for Heterogeneous FPGA Acceleration in the CloudIn Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) , 2025
- __INFOCOM__A Generic and Efficient Communication Framework for Message-level In-Network ComputingIn Proceedings of the IEEE International Conference on Computer Communications (INFOCOM) , 2025
- __ASPLOS__Design and Operation of Shared Machine Learning Clusters on CampusIn Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS) , 2025
- __EuroSys__Achieving Fairness Generalizability for Learning-based Congestion Control with JuryIn Proceedings of the 20th ACM European Conference on Computer Systems (EuroSys) , 2025
- __SIGCOMM__Fast, Scalable, and Accurate Rate Limiter for RDMA NICsIn Proceedings of the ACM Special Interest Group on Data Communication (SIGCOMM) , 2024
- __EuroSys__Astraea: Towards Fair and Efficient Learning-based Congestion ControlIn Proceedings of the 19th ACM European Conference on Computer Systems (EuroSys) , 2024
- ___NSDI___Accelerating Neural Recommendation Training with Embedding SchedulingIn Proceedings of the 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI) , 2024
- ___NSDI___Towards Domain-Specific Network Transport for Distributed DNN TrainingIn Proceedings of the 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI) , 2024
- __SIGMOD__Scalable and Efficient Full-Graph GNN Training for Large GraphsIn Proceedings of the ACM on Management of Data (SIGMOD) , 2023
- ___NSDI___SRNIC: A scalable architecture for RDMA NICsIn Proceedings of the 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI) , 2023
- ___ICNP___DGS: Communication-Efficient Graph Sampling for Distributed GNN TrainingIn Proceedings of the 30th IEEE International Conference on Network Protocols (ICNP) , 2022