HPC Systems EngineerNew$196K

Anduril Industries is a defense technology company with a mission to transform U. S.

What you'll do

• Work in a fast-paced, customer-focused environment supporting high-profile: operational and research requirements.

• Architect and deploy advanced GPU infrastructure, leading the design,: deployment, and lifecycle management of cutting-edge NVIDIA hardware including H100, H200, and B200/B300 systems.

• Ability to rack, stack, cable, and configure physical servers and multi-node GPU systems from end to end.

• Configure HPC and AI environments, including job schedulers (e.g., Slurm),: multi-user login environments, and cluster management software (e.g., Warewulf, NVIDIA Base Command, RunAI).

• Implement and fine-tune high-speed interconnects (e.g., NVLink, NVSwitch,: InfiniBand/NDR) crucial for large-scale distributed training.

• Configure and manage large-scale, high-performance storage platforms in the: multiple petabytes range, optimized for AI/ML data access patterns.

What they're looking for

• Install, configure, and maintain the application stack on HPC clusters,: including traditional simulation software (StarCCM+, Ansys, Matlab) and the core AI/ML software stack (NVIDIA drivers, CUDA, PyTorch, TensorFlow).
• Implement and manage GPU virtualization and sharing technologies, such as: Multi-Instance GPU (MIG), to maximize resource utilization across diverse workloads.
• Troubleshoot complex, system-wide issues related to application performance,: user access, compute nodes, storage, and job queueing services.
• Utilize NVIDIA Data Center GPU Manager (DCGM) and additional tools to: proactively monitor GPU health and performance, diagnosing and resolving training bottlenecks in collaboration with ML engineers.