Staff Network Engineer (AI Infrastructure)
About The Position
As a Staff Network Engineer at vCluster Labs, you will build automation systems that provision and configure the InfiniBand and high-speed Ethernet fabrics used by AI Clouds. You will begin in a small R&D group working alongside the CTO and grow into ownership of networking as it becomes a core part of the platform. You will automate NVIDIA switch and InfiniBand fabric provisioning, including integration with NVIDIA UFM; build systems that keep switch, VLAN, and VXLAN configuration aligned with platform intent; and develop resilient NAT gateway and Layer 4 load-balancing capabilities. You will define network behavior at scale, including routing limits, convergence, and failures under AI workloads, then validate your automation on real AI Cloud environments. You will also help set technical direction, decide what to build or integrate, write design and requirements documents, and establish networking standards across vCluster products. vCluster Labs builds infrastructure software for AI Cloud providers and AI factories. Its platform spans bare-metal provisioning, node lifecycle management, Kubernetes, Slurm, Ray, and inference clusters. The company describes itself as remote-first, with headquarters in San Francisco, and says it powers more than 100,000 GPUs and 1 million CPUs across over 50 AI clouds and Fortune 500 companies.
Requirements
- Experience building and shipping software that programmatically configures switches, routers, or load balancers, ideally for a product or internal platform used by other teams.
- Hands-on experience with NVIDIA UFM, NVIDIA Spectrum or Quantum switches, Cumulus Linux, or NVUE, including programming against their APIs.
- Strong knowledge of InfiniBand, VLAN, VXLAN and EVPN, and BGP, including their behavior and failure modes at production scale.
- Experience operating networks at production scale and resolving root causes with automation rather than runbooks.
- Ability to write production code in Go or Python.
- Use of version control and code review, and an approach that treats network configuration as code.
Nice To Haves
- Experience tuning fabrics for GPU training or inference workloads, including RoCEv2, congestion control, rail-optimized topologies, or Spectrum-X.
- Experience building network automation at a GPU cloud, hyperscaler, or large CDN.
- Experience building Layer 4 load balancers or NAT gateways, including with eBPF/XDP or DPDK.
- Contributions to networking or Kubernetes open-source projects such as SONiC, FRRouting, Cilium, or MetalLB.
Responsibilities
- Design and build systems that automatically provision and configure NVIDIA switches and InfiniBand fabrics, including integration with NVIDIA UFM.
- Build network configuration management and synchronization that keeps switch, VLAN, and VXLAN state consistent with platform intent, detects drift, and safely reconciles it in production.
- Build NAT gateway and Layer 4 load-balancing capabilities for tenant cluster connectivity, with failover designed in from the start.
- Define network behavior at scale, including BGP route table limits, convergence, failover, and responses to link, switch, or gateway failures under AI workloads.
- Operate networking automation in real AI Cloud environments, identify failures, and translate operational lessons into product behavior.
- Help determine what to build versus integrate with partner solutions, write design and requirements documents, and establish networking standards across vCluster products.
Benefits
- Competitive compensation package, including equity.
- Health, dental, vision, and life insurance for employees and eligible dependents; benefits vary by country.
- Flexible working schedule, with emphasis on results rather than fixed clock-in and clock-out times.
- Flexible work location and willingness to adjust the work environment as circumstances change.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
Education Level
No Education Listed
