renzhenghang at hotmail dot com
I work on GPU interconnect and high performance computing for AI workloads.
I gain my Ph.D. in the Hong Kong University of Science and Technology (HKUST).
Before joining HKUST, I received the bachelor degree from University of Science and Technology of China (USTC).
DevTech Engineer - NVIDIA 2025.12 - NOW
Efficient communication for AI through NVLink and RDMA.
Ph.D. in Computer Science and Engineering HKUST, 2020 - 2025
Bachelor in Computer Science USTC, 2016 - 2020
https://github.com/axio-project/FuseLink – Efficient GPU communication via multiple NICs.
https://github.com/axio-project - A high-perf, programmable, and extensible network system with DPU acceleration.
https://mixnet-project.github.io/ - A Runtime Reconfigurable Optical-Electrical Fabric for large-scale MoE Training
🎉 Jul 2026 PipeTune: An Efficient Performance Tuning Framework for Host Datapaths Accepted by NSDI 2027!
🎉 Jul 2026 Pyxis: Unlocking Multi-NIC Bandwidth for Incast Traffic in AI Clusters Accepted by NSDI 2027!
Enabling Efficient GPU Communication over Multiple NICs with FuseLink | USENIX
USENIX OSDI 2025
Accelerating Secure Collaborative Machine Learning with Protocol-Aware RDMA | USENIX
USENIX SECURITY 2024
APNET 2025
ASPLOS 2025
Revisiting RDMA Reliability for Lossy Fabrics | Proceedings of the ACM SIGCOMM 2025 Conference
SIGCOMM 2025