Member of Technical Staff, Compute Platform

Perimeter Compute
•$150,000 - $250,000•Remote

About The Position

Perimeter Compute is hiring a Member of Technical Staff to own the customer-facing compute products built on its bare-metal GPU infrastructure. The role begins with managed Kubernetes and extends to GPU virtual machines. Customers should experience the geographically distributed infrastructure as one cloud, with Perimeter responsible for failures such as node outages or loss of a site's connection. The successful candidate will own these products, including their service levels, monitoring, adoption, and direction. The company is building toward the criteria for the top tier of ClusterMAX, SemiAnalysis's public rating of GPU clouds. Success is measured by paying customers and the share of the fleet that is ready to rent at published service levels. Perimeter describes itself as a small, highly leveraged team where people own outcomes, work directly with customers, and make decisions independently. The company uses coding agents to write and review code, and expects its tools, APIs, documentation, and interfaces to work for both people and agents. The position is fully remote in the US. Visa sponsorship is not available.

Requirements

  • Experience building and operating production software that people depend on, including owning it when it fails.
  • Experience running production Kubernetes on bare metal, including provisioning, upgrades, and node operations.
  • Experience debugging production GPU failures, including drivers, Xid errors, and NCCL hangs.
  • Working knowledge of KVM/QEMU, KubeVirt, and GPU passthrough.
  • Experience owning a product that other developers paid for or depended on, including responsibility for its service levels.
  • Ability to get meaningful results from coding agents on infrastructure and identify which changes require human review.

Nice To Haves

  • Experience with hosted control planes using Gardener or k0smotron.
  • Experience with bare-metal provisioning using Ironic, Metal3, and Redfish.
  • Experience with network fabrics such as Netris or Hedgehog, and BlueField DPUs.
  • Experience with GPU partitioning, including MIG or NVLink partitions through Fabric Manager.
  • Experience with quotas and preemption using Kueue.

Responsibilities

  • Stand up managed Kubernetes on bare-metal infrastructure from scratch, including hosted control planes and the expected GPU and networking stack.
  • Build GPU virtual machines, beginning with whole-node offerings and adding smaller sizes as customers need them.
  • Create reserved, on-demand, and spot products from shared capacity, including quota and preemption rules and usage records for billing.
  • Build customer isolation into products to meet the standards set and tested by the security team.
  • Handle infrastructure failures, including moving work off failing nodes and keeping workloads running when a site loses its connection to the central control plane.
  • Make every operation accessible to agents as well as people through APIs, CLI, and agent tools.
  • Operate the products after launch: publish service levels, respond when they slip, and roll out Kubernetes and GPU driver upgrades without breaking customer workloads.
  • Talk with developers using the products and decide what to build next.

Benefits

  • Health insurance
  • Dental insurance
  • Vision insurance
  • Unlimited PTO
  • $500 home office stipend
  • Top-spec MacBook
  • Unlimited tokens
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service