AI Compute Lead

Ample Industries (Formerly Vortx); Santa Monica (CA); On-site; USD 180,000 - 280,000; Full time

Job summary

Ample Industries (Formerly Vortx) is seeking a hands-on leader to own the GPU stack across inference clusters on our power systems. You will design reference architectures, bring up racks and validate compute at each site, and oversee day-to-day operations.

You will drive cluster performance, manage uptime, thermal and power limits, and enable inference serving for API platform customers. Collaboration with the energy team will include co-designing load profiles and ride-through strategies.

Qualifications

• Experience deploying and operating GPU clusters at scale.
• Deep knowledge of inference serving and high-performance networking.
• Fluency in high-density power and cooling requirements.
• Experience serving external inference customers.
• Battery, microgrid or DC power experience is a plus.

Responsibilities

• Define reference architectures from NVL72-class racks down to smaller nodes.
• Rack, stack, power on and validate compute at each site.
• Run cluster operations: utilization, uptime, thermal and power limits.
• Enable inference serving for API platform customers, including scheduling and tenancy.
• Co-design load profiles and battery ride-through with the energy team.
• Own GPU, networking and integration vendors.

Skills

GPU clusters

Inference serving

InfiniBand/RoCE

Multi-tenancy

Power & cooling

External customers

Job description

You’ll own the GPU stack at every site: node design, bring-up and day-to-day operation of the inference clusters running on our power systems.

What you’ll do

• Define reference architectures from NVL72-class racks down to smaller nodes
• Rack, stack, power on and validate compute at each site
• Run cluster operations: utilization, uptime, thermal and power limits
• Enable inference serving for API platform customers, including scheduling and tenancy
• Co-design load profiles and battery ride-through with the energy team
• Own GPU, networking and integration vendors

What we’re looking for

• Hands-on experience deploying and operating GPU clusters at scale
• Deep knowledge of inference serving, InfiniBand/RoCE and multi-tenancy
• Working fluency in high-density power and cooling
• Experience serving external inference customers
• Battery, microgrid or DC power experience is a plus