Topia builds scalable, secure infrastructure for enterprise deployments that process billions of encrypted data connections annually. We are expanding our AI capabilities to power the next generation of our infrastructure-as-a-service.

This role is responsible for designing, deploying, and managing GPU infrastructure that powers Topia's AI and machine learning workloads. You'll work across on-premise clusters and cloud environments to build reliable, scalable compute infrastructure. The position combines hands-on systems work with architecture decisions that directly impact our AI capabilities.

• Design and deploy GPU clusters for training and inference workloads across on-premise and cloud environments
• Configure and optimize NVIDIA drivers, CUDA, container runtimes, and ML frameworks
• Build and maintain orchestration for GPU workloads using Kubernetes, Slurm, or similar schedulers
• Implement monitoring, alerting, and capacity planning for GPU resources
• Optimize infrastructure costs through efficient resource allocation and scheduling
• Troubleshoot hardware, driver, and software issues across the GPU stack
• Collaborate with ML engineers to understand workload requirements and optimize performance
• Document infrastructure architecture, runbooks, and operational procedures
• Evaluate and integrate new GPU hardware and cloud offerings

• 3–5 years of experience in infrastructure engineering, DevOps, or systems administration
• Hands-on experience with NVIDIA GPUs, CUDA, and container runtimes such as Docker or containerd
• Strong Linux systems administration skills
• Experience with GPU offerings from at least one major cloud provider such as AWS, GCP, or Azure
• Familiarity with Kubernetes or other container orchestration platforms
• Solid understanding of networking fundamentals including TCP/IP, DNS, and load balancing
• Scripting proficiency in Python, Bash, or similar languages
• Self-directed problem solver comfortable working autonomously

Nice to Have

• Experience with GPU cluster management tools such as Slurm, Ray, or DeepSpeed
• Knowledge of ML training workflows and frameworks such as PyTorch or TensorFlow
• Experience with InfiniBand or other high-speed networking technologies
• Familiarity with infrastructure-as-code tools such as Terraform or Pulumi
• Background in performance optimization and benchmarking
• Experience with multi-node distributed training setups

Competitive salary, stock options, health/dental/vision insurance, 401(k), and home office stipend.