AI Cluster Architect at Vultr

This role is designed for a heavyweight infrastructure expert with at least 7 years of experience in high-performance computing (HPC) or hyperscale environments. You are the ideal candidate if you possess a rare blend of networking expertise (InfiniBand/RoCE) and physical facility awareness. You should be comfortable calculating power envelopes and thermal loads just as easily as you design non-blocking network topologies for tens of thousands of GPUs. The compensation is competitive ($165k–$185k) for a fully remote U.S. role, but the real draw is the benefits package. Vultr offers 100% company-paid medical/dental premiums, a 4% 401(k) match with immediate vesting, and a unique longevity program including a one-month paid sabbatical every five years. As a mid-level architect in a company recently valued at $3.5 billion, you’ll have a front-row seat to scaling one of the world's largest private clouds. **You might be a good fit if you...** * Can design complex network fabrics like Clos or Dragonfly while accounting for switch radix and link speeds. * Have deep knowledge of GPU hardware (NVIDIA/AMD), NVLink, and PCIe node topologies. * Enjoy the "Tetris" of capacity planning—maximizing GPU density within strict power and cooling limits. * Are a documentation pro who can translate vendor SKUs into scalable architectural templates.