AI Cluster Architect

Vultr

2d•$200,000 - $220,000

About The Position

Vultr is seeking an AI Cluster Architect to create and refine large-scale GPU cluster architectures within strict power and infrastructure limits. The role focuses on power-aware design: determining the optimal GPU count for a fixed power envelope while accounting for compute nodes, storage, networking, cooling, and facility constraints. It requires experience with heterogeneous environments, multiple hardware generations, and end-user requirements. The architect evaluates how GPU SKUs, NICs, switches, and fabrics interact at scale, including their individual and aggregate power and thermal characteristics. The role assesses multi-plane, rail-optimized, and tiered fabric designs using technologies such as InfiniBand, RoCE, and SpectrumX, ensuring the network supports the intended GPU count within facility, switch-radix, and topology constraints. The architect balances customer requirements for compute, storage, and service density while maintaining acceptable GPU and fabric performance and maximizing usable GPUs within the power budget.

Requirements

At least 7 years of experience designing or building large-scale HPC, AI, or hyperscale GPU clusters.
Expert understanding of GPU and accelerator system design, including node topology, PCIe, NVLink, NVSwitch, ROCm, and NIC-to-GPU affinity considerations.
Strong familiarity with InfiniBand, RoCE, and SpectrumX networking, including multi-tier, multi-plane, Clos/dragonfly variants, and large-radix switch design.
Demonstrated experience modeling power draw and thermal characteristics of servers, GPUs, NICs, switches, optics, and storage systems.
Ability to design networks that maintain full non-blocking performance or intentionally use over- or under-subscription while understanding workload-performance impacts.
Proven ability to gather and analyze vendor SKU-level specifications and incorporate them into scalable cluster architectures.
Experience balancing customer-driven requirements for compute, storage, and service density with overall GPU count.
Strong documentation, communication, and cross-functional collaboration skills.

Responsibilities

Architect large-scale GPU clusters within fixed site power budgets, optimizing GPU density while reserving headroom for compute services, storage, and networking.
Model and validate power consumption across the full cluster bill of materials, including GPUs, CPUs, NICs, switches, fabric components, storage, and facility limits.
Evaluate tradeoffs among InfiniBand, RoCE, and SpectrumX networking architectures, including multi-plane, 2-tier/3-tier, and rail-optimized topologies.
Determine network scale limits based on switch radix, link speed, topology, and blocking requirements.
Gather, interpret, and maintain SKU-level power and thermal specifications for GPUs, NICs, switches, DPUs, storage, and server platforms.
Develop power-aware cluster configuration templates and capacity-planning models that scale across sites with varying constraints and support quick iteration and ideation.
Document architecture, design choices, tradeoff analyses, and operational considerations for deployment and lifecycle management.
Provide guidance on future-proofing designs to incorporate next-generation GPUs, NICs, or fabrics.
Collaborate with vendors on novel fabric architectures that enable large-scale cluster deployments of 100,000 or more GPUs.

Benefits

Company-paid employee medical, dental, and vision insurance premiums.
401(k) plan with a 100% match up to 4%, with immediate vesting.
Professional development reimbursement of $2,500 each year.
11 holidays, paid time off accrual, and a rollover plan.
Increased PTO at the 3-year and 10-year anniversaries.
One month of paid sabbatical every 5 years.
Annual anniversary bonus.
$500 remote-office setup stipend in the first year and $400 each following year.
Internet reimbursement up to $75 per month.
Gym membership reimbursement up to $50 per month.
Company-paid Wellable subscription.