Train your team

AI Infrastructure Engineer

Run GPU clusters that stay busy — the kernel, memory, fabric and driver layers under every training job.

6 courses 175 hours of instructionBadge: AI Infrastructure Engineer

Who it is for

Platform engineers responsible for expensive accelerators and the utilisation number they are judged on.

The sequence

  1. 1
    AIC-101 · Foundations of AI Compute & GPU Architecture 4 days
  2. 2
    AIC-200 · Linux for GPU Systems 4 days
  3. 3
    AIC-210 · RDMA & AI Cluster Networking 4 days
  4. 4
    AIC-220 · CUDA & HIP Programming 5 days
  5. 5
    AIC-230 · Distributed Training with PyTorch 4 days
  6. 6
    AIC-300 · Containers, Kubernetes & GPU Schedulers 4 days

Maps to the NVIDIA NCA-AIIO and NCP-AIN certification ladder.

Capstone

Specify, stand up and tune a multi-node GPU cluster, then show the utilisation curve that proves the money was spent well.

Start with the first course

Every course in the path can be taken on its own. Dates are on the course page.

AIC-101