Admit it, your breath quickens, you flush when someone whispers that naughty two letter acronym that has always been taboo to virtualize, that’s right, AI. Or maybe ML or DL is more your style, it doesn’t matter. In this session we fulfill your wildest nerd fantasies and demonstrate how to use Python and pyVmomi to construct virtual machines (VMs) for AI, ML, and DL workloads. We’ll show you those dirty little secrets you need to add multiple NVIDIA vGPUs to your VMs, so you can have huge VMs whenever it’s demanded from your favorite workload manager, be it Cloudera, Bright, or Slurm. The session will use pyVmomi to create VMs and control their configuration with configSpecs. It will explore what commands are needed to allow you to dynamically get the correct vGPU profile for your VM, that way you always get the right tool for the job.