← All roles

Qualification Engineer, Fluidstack Labs

Fluidstack· Development
Apply on company site
Austin, TX$202k - $241kFull-timePosted 3mo ago

About the role

Role focused on bringing up and qualifying first-sample accelerator hardware and associated infrastructure for AI compute at Fluidstack Labs, including hardware, firmware, network, liquid cooling, and runbook development to production-ready standards.

Responsibilities

  • - Bring up first-sample accelerator platforms (NVIDIA, AMD, custom accelerators) end to end: rack integration, liquid cooling commissioning, firmware baseline establishment, network connectivity, and software stack validation at rack densities up to and beyond 120 kW.
  • - Validate network platforms across Broadcom Tomahawk, Broadcom Jericho, and NVIDIA Spectrum silicon, drive Keysight Ixia traffic generation for RFC 2544/2889 benchmarking, line-rate stress, and protocol correctness.
  • - Verify the optical layer with EXFO test equipment, covering BER characterization and power budget analysis across the link inventory a qualification depends on.
  • - Qualify CDUs and liquid cooling across nVent, CoolIT, and Vertiv platforms: commissioning procedures, BMS telemetry integration, leak detection validation, coolant chemistry verification, and the operational runbooks that fall out of each pass.
  • - Exercise the full power oversubscription response chain: graceful and forced shutdown paths, power-cap and p-state levers via BMC, ATS transfer scenarios, breaker-trip detection, rPDU commissioning with outlet-level telemetry, and repeatable power-virus stress harnesses against accelerator hardware.
  • - Co-develop qualification matrices with hardware partners, evaluate converged local-NVMe storage platforms (Weka, Hammerspace, VAST Data), and turn results into runbooks production teams inherit.

Qualifications

  • - You've personally brought up servers, accelerators, or switches from first power-on: racking, cabling, firmware, first boot, and everything that goes wrong in between.
  • - You script your test harnesses and automation rather than clicking through runs, so your results are repeatable and your throughput compounds.
  • - You're deep in Linux and manage hardware at the BMC level: IPMI, Redfish, sensor telemetry, and power and boot control from the command line.
  • - You document methodically, producing test reports and runbooks that someone else can execute without you in the room.
  • - You've worked hands-on with liquid-cooled, high-power hardware and treat its safety procedures as part of the craft.
  • - You debug across hardware, firmware, and software layers yourself, without waiting for someone else to isolate the fault.
  • - Bonus: RDMA/RoCEv2 fabrics. Optics testing depth. Kubernetes-based provisioning stacks. Power and electrical instrumentation experience.

Skills

Required:Server and accelerator bring-upTest automation scriptingLinux system administration (BMC level)Documentation & runbook authoringLiquid cooling & high-power hardware safetyCross-layer debugging (HW/FW/SW)
Nice to have:Bonus: RDMA/RoCEv2 fabrics
Last updated Sep 15, 2026. Interested? Apply on the company site.
Fluidstack logo

About Fluidstack

Fluidstack builds civilization-scale infrastructure for AI

Fluidstack builds civilization-scale infrastructure for AI, delivering gigawatts of compute in 6 months. We're leading the deployment of Anthropic's $50B compute build out.

AI infrastructurecompute centershardwaresoftwarepowerdata centersAnthropicfrontier compute
Company profileWebsite

More open roles at Fluidstack