Staff Network Engineer (AI Infrastructure)

As a Staff Network Engineer at vCluster Labs , you aren't just configuring switches; you are building the automation logic that configures them. AI Clouds run on InfiniBand and high-speed Ethernet fabrics that most teams still wire up by hand, one customer at a time. You will build the automation that turns that fabric into repeatable provisioning logic. You start in a small R&D group alongside our CTO and grow into owning this area as it becomes a core part of our platform. As a Staff Network Engineer, your role will include: Fabric automation: You will design and build the systems that automatically provision and configure NVIDIA switches and InfiniBand fabrics, including integration with NVIDIA UFM. Manual, per-customer switch setup is the thing you are replacing. Network config management and sync: You will build the control loop that keeps switch, VLAN, and VXLAN state consistent with what our platform intends, detects drift, and reconciles it safely in production. NAT gateway and L4 load balancing: You will build the gateway and load balancing layer that tenant clusters depend on for connectivity, with failover designed in from day one rather than bolted on later. Scale and failure engineering: You will define how our network layer behaves at large scale: BGP route table limits, convergence, failover, and what happens when a link, switch, or gateway dies under real AI workloads. Customer Zero for AI networking: You will run your automation on real AI Cloud environments, bring back what breaks, and turn hard-won operational lessons into product behavior our customers never have to think about. Technical direction: You will help decide what we build versus integrate with partner solutions, write the design and requirements docs, and set the standards for how networking works across vCluster products. This role could be a fit for you if you bring: You have built network automation, not just operated networks: You have shipped software that configures switches, routers, or load balancers programmatically, ideally as part of a product or internal platform other teams relied on. NVIDIA networking experience: Hands-on work with NVIDIA UFM, NVIDIA Spectrum or Quantum switches, Cumulus Linux, or NVUE, including programming against their APIs. Deep fabric fundamentals: Strong command of InfiniBand, VLAN, VXLAN and EVPN, and BGP, including how they behave and fail at production scale. Production scale scars: You have run networks at real production scale, seen them break, and helped fix the root cause with automation rather than runbooks. Software engineering ability: You write production code in Go or Python, use version control and code review as a matter of course, and think of network config as code. Bonus points for: AI networking depth: Experience tuning fabrics for GPU training or inference workloads, such as RoCEv2, congestion control, rail-optimized topologies, or Spectrum-X. AI Cloud or hyperscaler background: You helped build network automation at a GPU cloud, hyperscaler, or large CDN. Load balancer and gateway internals: Experience building L4 load balancers or NAT gateways, including with eBPF/XDP or DPDK. Open source: Contributions to networking or Kubernetes projects such as SONiC, FRRouting, Cilium, or MetalLB. About vCluster Labs We're the #1 platform for AI infrastructure, trusted by the world's fastest-growing AI cloud builders. We're a venture-backed startup that's raised over $28M from top-tier investors including Khosla Ventures (first investor in OpenAI, GitLab, Stripe, and DoorDash), and we're in a hyper-growth phase looking for motivated people to join our team. Our headquarters are in San Francisco (Salesforce Tower), but our team is distributed around the globe with a remote-first culture. We give AI Cloud providers and AI factories a hyperscaler-like experience