Staff, Platform Engineer AI Compute (6034)
Core
Build and maintain the software and configuration for a shared GPU compute platform (AI Factory) to support machine-learning engineers in training and evaluating models.
Role type
Senior Staff Platform Engineer (AI Compute Infrastructure)
Builds
Shared GPU compute platform, Kubernetes clusters, and deployment automation tools for ML workloads
Domain
Defense technology, Artificial Intelligence, Distributed Systems
Deliverable
infrastructure
Required skills
Kubernetes, Python, Go, Linux, Networking, Storage, Infrastructure as Code (Terraform, Helm), GPU workload scheduling, Troubleshooting
Preferred skills
Run:ai, Ray, PyTorch, Air-gapped platform deployment, Open-source infrastructure
Responsibilities
Build and maintain software for running GPU workloads on Kubernetes; Design quotas, scheduling priorities, and isolation policies; Automate deployment and upgrades of GPU services; Investigate and fix scheduling, configuration, and networking issues; Build diagnostics for queue times and GPU utilization; Benchmark workloads to recommend GPU capacity expansion.