An AKS cluster must run ML inference workloads requiring NVIDIA GPUs. Which step is required?

1. A Enable GPU flag on any node pool
2. ✓ B Use node pools with GPU VM SKUs (NC/ND/NV) and install nvidia-device-plugin DaemonSet (correct answer)
3. C GPUs are automatically detected and configured
4. D Use ACI burst for GPU workloads

Shortcut: GPU workloads require node pools using GPU VM SKUs. AKS can automatically install the NVIDIA device plugin when using a GPU node pool with the appropriate SKU.