Senior GPU Cluster Network Engineer
sciforium · San Francisco
Job description
About the role
Sciforium is building next‑generation AI infrastructure and needs a senior engineer to own the networking stack of its large‑scale GPU clusters. You will design, bring up, and operate the full fabric—from InfiniBand/RoCE compute fabrics to perimeter firewalls and cloud interconnects.
Key responsibilities
- Architect greenfield network designs for new GPU clusters, including compute, storage, in‑band, out‑of‑band, and perimeter topologies.
- Select and justify RDMA fabrics (InfiniBand or lossless Ethernet/RoCE v2) for training and inference workloads.
- Specify switches, NICs, optics, and cabling; produce port maps and cable schedules for hardware operations.
- Define IP addressing, VLAN/VRF segmentation, and BGP/EVPN‑VXLAN underlay‑overlay architecture.
- Configure and validate RoCE v2 at scale (PFC, ECN/DCQCN, QoS) and manage InfiniBand subnet managers, adaptive routing, and SHARP.
- Run performance benchmarks (perftest, nccl‑tests) and troubleshoot fabric‑level issues.
- Maintain routing protocols (BGP, OSPF, ECMP) and switch operating systems across multiple vendors.
- Implement network‑as‑code pipelines using NetBox, Ansible/Nornir/NAPALM, and Git‑based CI.
- Design dark‑fiber/DWDM inter‑site links and hybrid‑cloud connectivity (AWS Direct Connect, Azure ExpressRoute, GCP Interconnect).
Required profile
- 7+ years designing and operating production data‑center networks, with experience on at least one large‑scale HPC/AI cluster fabric.
- Deep expertise in routing (BGP, EVPN‑VXLAN, OSPF) and RDMA (RoCE v2 tuning or InfiniBand management).
- Hands‑on experience with 100‑800 Gbps optics, high‑radix switches, and Clos/rail‑optimized topologies.
- Proficiency in network automation (Python, Ansible, Git‑based workflows).
- Knowledge of host‑side RDMA stacks (MOFED, DOCA, GPUDirect) and how NCCL/RCCL map onto the fabric.
Required skills
- InfiniBand, RoCE v2, BGP, EVPN‑VXLAN, OSPF, ECMP, VRF
- Arista EOS, NVIDIA Cumulus, Cisco NX‑OS, Junos, SONiC
- Python, Ansible, Nornir, NAPALM, Git, CI pipelines
- NetBox, Prometheus, Grafana, sFlow, gNMI
- AWS Direct Connect, Azure ExpressRoute, GCP Interconnect
What we offer
- Medical, dental, and vision insurance
- 401k plan
- Daily lunch, snacks, and beverages
- Flexible time off
- Competitive salary and equity
Questions fréquentes
Why are you reporting this job?
Explore further
Salaries, guides and searches in the United States.
Salaries by job title
Apply in 30 seconds
Enter your email to apply. An account will be created automatically.
By continuing, you accept our terms of use.
Already have an account? Login
Published 10 hours ago
Expires 1 month from now
5 views · 0 interested
Boost your chances
Upload your CV — we will match you with relevant openings.
Analyzing your CV...
sciforium
San Francisco
Related job offers
-
Strategist, Agent Development - Tech, Media & Telecom
sierra San Francisco -
Network Engineer – Secure AI Data Center
security-level-5 San Francisco -
Network Engineer – Global Office Networks
cognition San Francisco -
Principal Network Security Architect
cerebras United States and Canada -
Danish-Speaking Customer Support Associate – Telecommunications
Employeur non precise Sofia