Software Engineer (Infrastructure, Go)
Posted Updated
- We are looking for an experienced Full-Stack Software Engineer to design and implement the Infrastructure Services that power our GPU-as-a-Service platform
- You will build the control plane that turns high-level API calls into real infrastructure actions — enrolling bare-metal servers, provisioning them, and assembling them into multi-tenant Kubernetes clusters on high-performance hardware
- You will own the full lifecycle of the infrastructure-level services—from the Server and MachineType APIs down to the provisioning workflows and reconciliation loops that keep the platform’s view of hardware consistent with physical reality
- Infrastructure API Design: Design, build, and maintain the versioned REST and gRPC APIs for bare-metal server lifecycle, MachineType definitions, and cluster CRUD operations
- Provisioning & Workflow Development: Develop the asynchronous workflows that drive server enrollment, inspection, OS provisioning, and cluster bring-up, exposing durable status to callers
- System Reliability: Design the error-handling models, idempotency guarantees, and reconciliation loops necessary to manage long-running provisioning operations reliably
- Kubernetes Knowledge: Deep understanding of Kubernetes primitives and controller/reconciler patterns. You will be interacting with systems like k0rdent, Metal3, and Cluster API to translate high-level API calls into infrastructure actions
- Bare-Metal Provisioning: Hands-on experience with bare-metal provisioning flows — BMC/Redfish, PXE/iPXE, image management, and hardware inspection
- API Development: Strong experience designing RESTful APIs or gRPC services. You understand API versioning and gateway patterns
- Proficiency in Go (preferred for backend/Kubernetes ecosystem)
- Asynchronous Systems: Experience building workflow-driven or event-driven systems (e.g., Temporal) where operations are long-running and state must remain consistent across retries and failures
- State Reconciliation: Experience building informers or reconciliation bridges that keep an external datastore consistent with Kubernetes resource state
- Multi-Tenancy: Experience building platforms where strict data and network isolation between tenants is required
- Infrastructure-as-Code: Familiarity with Terraform/OpenTofu and GitOps-driven configuration (ArgoCD or Flux)
- Hardware Domain: Familiarity with GPU server hardware, DPUs/NICs, and high-performance datacenter fabrics