About client
Client builds and operates GPU cloud infrastructure at the scale modern AI demands. We run dense accelerated-compute clusters on bare metal, powered by NVIDIA and AMD accelerators and K8s, stitched together with high-performance Ethernet and InfiniBand fabrics, and we deliver them to customers as reliable, secure, high-throughput platforms. We are scaling rapidly across new data center sites, and the infrastructure team sits at the center of that growth, building the systems that turn racks of hardware into a product customers depend on.
The role
We are hiring a Principal Infrastructure Engineer to build the foundation of our greenfield v2 cloud platform: the bare-metal provisioning and control-plane substrate that turn new racks of GPU hardware into reliable, secure, customer-ready capacity.This is a build-focused, hands-on role and the most senior individual contributor on the platform build. You will own the recommendation and the call on the infrastructure layer (we’re currently in POC across different vendors) and the build-vs-buy decision that goes with it, with our Cloud Platform Architect pressure-testing it. You will drive integration across the other domains and define how infrastructure is templated, multi-tenanted, and handed to customers. You will define the architecture and the standards the rest of the team builds to, while still being deep in the code yourself. Day-to-day operations are not your focus, getting the platform designed and built right is.
You will work at the center of new cluster and new-site buildouts, partnering closely with our network, SRE, and security engineers. They own the fabric and day-to-day operations; you own getting the platform built right. Because the systems you design become the foundation customer GPU workloads run on, getting the architecture and automation right the first time matters more here than almost anywhere.
What you’ll own
• Platform architecture and technical direction: own the architecture of the cloud platform control plane and orchestration layer for new GPU clusters, making the foundational decisions and setting the technical standards the rest of the platform is built on.
• Bare-metal provisioning automation: build the automated pipeline that takes a node from rack-and-stack handoff to a fully configured, real acceptance gates before a node is marked customer-ready, firmware to the qualified recipe, GPU burn-in, NCCL/RCC runs, and fabric checks, plus a golden-image build-and-qualify pipeline in CI repeatedly across sites and at scale.
• Infrastructure as code and automation: treat the platform as code from day one. Build the Terraform and Ansible that provision, configure, and validate the platform, so the build is reproducible rather than hand-assembled.
• Control-plane substrate: build the control plane that turns provisioned bare metal into cluster-ready capacity, and own the clean handoff to the Kubernetes layer. You own tenant isolation at the network and hardware layer; the Kubernetes and namespace isolation model belongs to the Principal Kubernetes Platform Engineer.
• DPU and SmartNIC integration: design how we leverage BlueField DPUs in the platform, offload, provisioning, isolation, and the SuperNIC/DPU-mode configurations that let us get the most out of the hardware.
• Platform reliability by design. Build observability, guardrails, and progressive-rollout mechanisms into the platform from the start. As you build it and design to our SOC 2, ISO 27001, and HIPAA obligation.
• Security Built-in: instrument the platform as you build it and design to our SOC 2, ISO 27001, and HIPAA obligations. Vault-based secrets model from the start. BIOS and system prep for trusted domain execution, secure boot, and backend-fabric hardening.
• Confidential-compute-ready platform (growth area): Design the tenancy and scheduling so it fits our confidential-compute roadmap: whole-GPU per tenant, and TDX with
