腾讯云-高性能网络高级后台研发工程师(RDMA方向)
Core
Design high-performance network architectures for AI clusters and develop low-latency communication engines for large-scale distributed model training.
Role type
Senior IC backend engineer (high-performance networking, RDMA)
Builds
AI cluster networking infrastructure, NCCL/UCX communication libraries, DPU hardware offloading engines
Domain
Cloud computing, AI infrastructure, high-performance networking
Deliverable
production ML models
Required skills
C/C++/Python, Linux TCP/IP protocol stack, RDMA/RoCE, DPDK, network congestion control, traffic scheduling, DPU programmable data plane, large-scale GPU cluster debugging
Preferred skills
Public cloud business scenarios, cloud migration solutions, industry technology trends
Technologies
RDMA, RoCE, NCCL, UCX, DPDK, VSWITCH, DPU
Responsibilities
Design AI cluster network architectures, optimize collective communication libraries, develop hardware offloading engines, resolve network performance bottlenecks in large GPU clusters
Seniority
Senior, hands-on IC
