Systems/Network Engineer - High-Performance Compute GPU Infrastructure

Role Overview As a Systems/Network Engineer, you will be responsible for architecting, deploying, and maintaining GPU-based compute infrastructure. You will work on bare-metal systems, high-speed networks, and hybrid cloud integrations to ensure maximum performance, reliability, and scalability. This role is primarily remote but may occasionally require on-site support for hardware installations or emergency maintenance. Key Responsibilities: System Optimization

• Configure and optimize bare-metal servers, including Linux OS, NVIDIA/AMD GPU drivers, and system libraries.
• Fine-tune NUMA settings, CPU-GPU affinity, and storage I/O for peak performance.
• Benchmark and tune HPC systems for specific workloads, ensuring sustained high performance.

• Deploy and manage GPU clusters using job orchestration tools like Kubernetes, Slurm, or similar platforms.
• Monitor GPU utilization, thermals, and overall system health using tools like NVIDIA DCGM, ROCm, and Prometheus/Grafana.

• Design and maintain high-speed networking solutions (e.g., NVLink, InfiniBand, RDMA) for distributed GPU systems.
• Optimize data transfer between nodes and reduce latency in cluster communication.

• Manage and configure storage solutions such as NVMe, SSD arrays, Ceph, or Lustre for high-throughput workloads.

• Automate system deployment, updates, and monitoring using tools like Ansible, Terraform, or Python scripts.

• Implement secure access controls, firewalls, and VPNs to protect GPU resources and user data.
• Ensure compliance with security best practices for HPC environments.

• Manage integrations between on-premise GPU clusters and cloud platforms (e.g., AWS, GCP, Azure).
• Build and maintain hybrid HPC setups for seamless scalability.

• Work on power, cooling, and rack design for HPC setups, ensuring reliable and efficient operations.
• Deploy and maintain systems in on-premise or hybrid cloud data center environments.

• Strong experience with Linux (CentOS, Ubuntu, RHEL) and system-level configuration.
• Expertise in managing NVIDIA GPU ecosystems (CUDA, NVLink, NVIDIA drivers).
• Familiarity with AMD ROCm, HIP, or OpenCL for AMD GPUs.
• Knowledge of high-speed networking protocols (InfiniBand, RDMA, Ethernet).
• Proficiency in scripting and automation (Python, Bash, Ansible, Terraform).
• Experience with job orchestration tools like Kubernetes or Slurm.
• Familiarity with containerization (Docker, NVIDIA Docker, Singularity).
• Understanding of storage technologies, including NVMe and parallel file systems.

• Strong analytical and problem-solving skills.
• Ability to work independently and as part of a remote team.
• Excellent communication skills for cross-team collaboration.

• Experience with hybrid cloud setups, including AWS Outposts, Azure Stack, or GCP Anthos.
• Hands-on experience with hardware management tools like IPMI/BMC for remote server management.
• Familiarity with emerging accelerators (e.g., SambaNova, Cerebras, Graphcore).

• Competitive salary and benefits package.
• Work with a talented and collaborative team of engineers.
• Opportunities to work on cutting-edge GPU and HPC projects.
• A flexible and dynamic startup environment where you can grow and innovate.
• Opportunities for professional development and continuous learning.