Staff Infrastructure Engineer, Networking (Groq Metal)
About The Position
This role is for a Sr Staff/Staff Infrastructure Engineer who will own the bare-metal layer of the platform, focusing on how thousands of physical servers across multiple data centers become healthy, secure, tenant-assignable Kubernetes nodes and maintain that state. This position is part of the Platform Engineering team and serves as the technical lead for all aspects below the Kubernetes API, including server provisioning, OS lifecycle, datacenter networking fabric, BMC/firmware management, and integration with modern DPUs. The role requires the individual to be based in the Dallas, San Francisco, or New York City area, with an initial remote work flexibility that will transition to onsite once the local Groq office is established.
Requirements
- 8+ years in production infrastructure, with at least 3 years owning bare-metal lifecycle at scale.
- Strong understanding of datacenter networking fundamentals: BGP, leaf/spine, MTU, congestion control, fabric debugging.
- Production Kubernetes operations experience, including cluster bring-up, CNI internals, and fleet upgrades.
- Proficiency with infrastructure-as-code tools (Terraform, Helm) and GitOps.
- Experience being on-call for a fleet of physical machines.
Nice To Haves
- Experience with NVIDIA GPU clusters at scale (H100 / GH200 / GB200 / NVL72), including driver stack, NCCL, NVLink, and GPU Operator.
- Familiarity with DPUs (NVIDIA BlueField, AMD Pensando) and offload patterns.
- Experience with Spectrum-X / RoCE / InfiniBand fabrics for AI workloads.
- Background in hyperscaler infrastructure, AI cloud (CoreWeave / Crusoe / Lambda / Together), HFT, or HPC.
Responsibilities
- Bare-metal lifecycle at scale, including automated server bring-up (PXE/iPXE, Metal3, Ironic, kubeadm), OS image lifecycle, BIOS/firmware/BMC fleet management, and hardware health telemetry and remediation.
- Datacenter networking fabric management, including leaf/spine architecture, BGP, ECMP, RoCE, congestion control, tenant traffic isolation, IPAM, and switch automation. This involves partnering with networking architects on Spectrum-X-class fabrics for multi-tenant AI workloads.
- Managing the Kubernetes substrate, specifically the host clusters that run the multi-tenant control plane (Cilium CNI, Karpenter for capacity, GitOps via Flux/Argo, virtual cluster integration), ensuring a stable foundation for Kubernetes application developers.
- Leading new datacenter bring-up, transforming the process from physical setup to production-ready capacity through repeatable, automated workflows.
- Overseeing Day-2 operations, collaborating with DC field staff on hardware faults, maintenance windows, and on-call response, and representing the physical layer in platform design discussions.
Benefits
- Competitive compensation through Total Cash philosophy (inclusive of potential bonus value).
- Quarterly bonus as an interim alternative to equity.
- Plans to introduce long-term incentive program participation.
- Robust suite of benefits.
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
