Hey!
I have experience working with Google Cloud and gcloud command. I have setup multiple models and I have trained it on Google Cloud ML Engine. The basic workflow would be setting up development environment locally and write the code and test it. Once it is done, we can push our code files to ML Engine by gcloud ml-engine command and our training, validation and testing data can be put on Cloud Buckets. We can even check the logs to know how our training process is going on etc.