HPC / GPU Cluster Architect

About the role

sfcompute is building a liquid market for GPU offtake, enabling companies to mitigate the financial risk of large‑scale compute investments. As a HPC/GPU Cluster Architect you will join a small, ambitious infrastructure team responsible for designing, deploying, and operating high‑performance compute clusters worldwide.

Key responsibilities

• Participate in an on‑call rotation to ensure cluster availability.
• Deploy new GPU compute environments and scale existing clusters.
• Diagnose and resolve performance, reliability, and hardware issues across the stack.
• Automate provisioning, monitoring, and remediation using infrastructure‑as‑code practices.
• Create and maintain operational documentation and runbooks.
• Mentor junior engineers and help shape team culture.

Required profile

• 5+ years of hands‑on experience designing, architecting, and scaling at least one production HPC or GPU compute cluster.
• Deep understanding of server hardware fundamentals (GPUs, NICs, PCIe, memory, thermals, power).
• Strong debugging skills across hardware, OS, drivers, and networking layers.
• Passion for automating fleet operations and infrastructure‑as‑code.
• Ability to produce clear operational documentation and runbooks.
• Willingness to work onsite in San Francisco 3–4 days per week and travel domestically as needed.

Required skills

• GPU cluster architecture
• Server hardware (GPUs, NICs, PCIe, memory, thermals, power)
• Linux operating system
• Infrastructure‑as‑code automation
• Monitoring and remediation tools

What we offer

• Opportunity to shape a pioneering market for GPU compute.
• Early‑stage, high‑impact environment with rapid growth.
• Collaborative team culture and mentorship opportunities.