Principal Software Engineer, GPU Compute

📋 Description

• Lead GPU compute for the Compute team across Kubernetes, Networking, and Cloud.
• Own the GPU host lifecycle: driver, firmware, health, telemetry, remediation.
• Architect GPU capacity exposure to compute platforms and Kubernetes.
• Drive fleet reliability and automated repair of GPUs.
• Evaluate onboard new GPU/AI accelerators; NVLink/InfiniBand/RoCE.
• Set standards, tooling, and APIs for safe GPU compute use.

🎯 Requirements

• 10+ years building/operating large-scale distributed systems.
• Deep GPU expertise: provisioning, driver/firmware, health, reliability.
• Proven compute expert; scaled GPU infrastructure others depend on.
• Strong Go or similar language proficiency.
• Experience with GPU/AI workloads; CUDA, scheduling, NVLink, InfiniBand, RoCE.
• Familiarity with Kubernetes for GPU workloads; bare-metal concepts.
• Anchor expert with leadership to elevate engineers around you.

🎁 Benefits

• Onsite schedule: in-office Tue, Wed, Thu; optional Mon/Fri.
• Equal employment opportunities for all applicants.
• Reasonable accommodations during the recruiting process.