@pfhllnts I hope you don't just dockerize your ML workloads but at least have kubeflow as well.