Senior AI Compute Architect — Scalable GPU Clusters

Lambda; Bellevue (WA); On-site; USD 314,000 - 465,000; Full time

Job summary

Lambda, The Superintelligence Cloud, is a leader in AI cloud infrastructure serving tens of thousands of customers. Our customers range from AI researchers to enterprises and hyperscalers. Lambda's mission is to make compute as ubiquitous as electricity and give everyone the power of superintelligence.

One person, one GPU. You’ll architect scalable compute platforms optimized for AI/ML and high-throughput workloads, define standards, evaluate CPU/GPU/accelerator technologies, and collaborate

Qualifications

• 7+ years of experience architecting large-scale GPU HPC or cloud compute platforms.
• Deep knowledge of CPU/GPU architectures and accelerator topologies.
• Experience with high-bandwidth, low-latency interconnects (NVLink, InfiniBand, RoCE).
• Ability to drive architecture decisions across teams.

Responsibilities

• Architect scalable compute platforms for AI/ML workloads and high-throughput tasks.
• Define standards and design patterns for compute infrastructure.
• Evaluate CPU/GPU/accelerator technologies and tradeoffs for density and cost.
• Collaborate with product and engineering to map workloads to platform capabilities.
• Lead platform introductions and mentor engineers on performance tuning.

Skills

GPU HPC architecture

CPU/GPU architectures

High-bandwidth fabrics

Performance tuning

Cross-functional leadership

Ownership mindset

Tools

Slurm

Kubernetes

GPU virtualization

Job description

Lambda, The Superintelligence Cloud, is a leader in AI cloud infrastructure serving tens of thousands of customers. Our customers range from AI researchers to enterprises and hyperscalers. Lambda's mission is to make compute as ubiquitous as electricity and give everyone the power of superintelligence.

One person, one GPU. You’ll architect scalable compute platforms optimized for AI/ML and high-throughput workloads, define standards, evaluate CPU/GPU/accelerator technologies, and collaborate