SRE - Compute & Hyperscale GPU Fleet Reliability

Fluidstack; New York (NY); On-site; USD 175,000 - 300,000; Full time

Benefits offered by this job

Health, dental, and vision insurance

Equity participation

Retirement plan

Generous PTO

Job summary

Fluidstack is seeking a Production Engineer to own compute fleet health end-to-end, build the observability and automation that scales GPU infrastructure, and drive reliability across Kubernetes-managed and bare-metal environments. You will define failure modes, implement triage automation, and own the GPU qualification and firmware tooling.

The role emphasizes end-to-end ownership, rapid incident response, and fluency with AI tooling and modern automation.

Qualifications

• Toil treated as a bug; repairs workflow are not for one-off efforts.
• Experience reasoning about hardware failure modes at firmware or silicon level.
• Ability to navigate ambiguity and map complex systems quickly.
• Rapid learning in unfamiliar domains with strong competence.
• Ability to handle incidents, write postmortems, and fix systemic causes.
• Familiar with AI tooling, LLM APIs, MCP servers, agentic frameworks.
• Experience shipping production automation used by other teams.
• Bonus: familiarity with BMC/Redfish/IPMI, GPU qualification, and orchestration tools.

Responsibilities

• Own compute fleet health end-to-end with observability and fleet-scale tooling.
• Turn deployment and repair into scalable pipelines; automate triage, parts management, and return-to-service.
• Design and expand the GPU qualification platform for new generations.
• Own Redfish and BMC tooling for firmware telemetry and fleet-access layers.
• Ensure reliability, scalability, and operation of the GPU compute fleet at scale.

Skills

Toil-as-bug

Hardware intuition

Ambiguity tolerance

Fast learner

Pager incident handling

AI tooling fluency

Production automation

RMA automation (bonus)

Tools

Go

Python

Job description

Fluidstack is seeking a Production Engineer to own compute fleet health end-to-end, build the observability and automation that scales GPU infrastructure, and drive reliability across Kubernetes-managed and bare-metal environments. You will define failure modes, implement triage automation, and own the GPU qualification and firmware tooling.

The role emphasizes end-to-end ownership, rapid incident response, and fluency with AI tooling and modern automation.