Title: Linux Admin Location: Remote. Key Responsibilities
Infrastructure Management: Provision, deploy, and maintain scalable, secure, and
high-availability cloud infrastructure on platforms such as Cloud to support
AI workloads.
System Management: Administer and maintain Linux-based servers and clusters
optimized for GPU compute workloads, ensuring high availability and performance.
GPU Infrastructure: Configure, monitor, and troubleshoot GPU hardware (e.g., NVIDIA
GPUs) and related software stacks (e.g., CUDA, cuDNN) for optimal performance in
AI/ML and HPC applications.
Troubleshooting: Diagnose and resolve hardware and software issues related to GPU
compute nodes and performance issues in GPU clusters.
High-Speed Interconnects: Implement and manage high-speed networking
technologies like RDMA over Converged Ethernet (RoCE) to support low-latency,
high-bandwidth communication for GPU workloads.
CI/CD Pipelines: Build and optimize continuous integration and deployment (CI/CD)
pipelines for testing GPU-based servers and managing deployments using tools like
GitHub Actions.
Monitoring & Performance: Set up and maintain monitoring, logging, and alerting
systems (e.g., Prometheus, Victoria Metrics, Grafana) to track system performance,
GPU utilization, resource bottlenecks, and uptime of GPU resources.
Security and Compliance: Implement network security measures, including firewalls,
VLANs, VPNs, and intrusion detection systems, to protect the GPU compute
environment and comply with standards like SOC 2 or ISO 27001.
Required Qualifications
Experience: 3+ years of experience in DevOps, Site Reliability Engineering (SRE), or
cloud infrastructure management, with at least 1 year working on GPU-based compute
environments in the cloud.
Linux Administration: Strong knowledge of Linux system administration for managing
network services and tools in a GPU compute environment.
High-Speed Interconnects: Experience with high-performance networking technologies
like RoCE, or 100GbE Ethernet in compute-intensive environments.
GPU-Specific Networking: Proficiency with NVIDIA GPU networking technologies,
such as Mellanox ConnectX adapters, and configuring Netplan to support their drivers
and firmware.
Cloud Platforms: Hands-on experience with at least one major cloud provider (AWS,
Azure, Google Cloud Platform).
Networking & Security: Knowledge of networking concepts (VPC, subnets) and
security best practices (IAM, encryption, firewall configurations).
|