Senior Network Engineer – GPU Cluster Networking
Core
Architect, deploy, and operate high-performance backend networks for large-scale AMD Instinct GPU clusters to support AI training, inference, and HPC workloads.
Role type
Senior IC network engineer (GPU cluster fabrics)
Builds
High-speed Ethernet and RoCEv2 network fabrics for 10,000+ GPU clusters
Domain
Data center networking, AI/HPC infrastructure, RDMA
Deliverable
production ML models
Required skills
RoCEv2, RDMA, high-speed Ethernet (100/200/400 GbE), leaf-spine/Clos/fat-tree architectures, BGP, ECMP, EVPN, VXLAN, switch buffering, PFC, DCQCN, Linux networking, PCIe/NUMA topology, capacity planning, incident response
Preferred skills
Juniper Junos OS, AMD Instinct accelerators, ROCm, RCCL, AMD Pensando NICs/DPUs, Prometheus/Grafana observability, large language model training networks
Technologies
RoCEv2, 100/200/400 GbE, BGP, ECMP, EVPN, VXLAN, PFC, DCQCN, Linux, Juniper Junos, ROCm, RCCL, Prometheus, Grafana
Responsibilities
Design network fabrics for 10,000+ GPU clusters; configure and tune lossless RoCEv2 environments; perform topology modeling and capacity planning; lead production incident response and root-cause analysis; optimize end-to-end communication performance across GPUs, NICs, and switches; execute network expansions and fabric migrations
Seniority
Senior, hands-on IC
