How do I make my models available to my applications?

We introduced this concept in Chapter 1, and we talked about the various things you would need to do to host a model on your own, such as setting up all of the required infrastructure, including load balancers, routers, switches, cables, servers, and storage, among other things, and then managing all of that infrastructure on an ongoing basis. This would require a lot of your time and resources.

Luckily, all of that stuff was in the old days, and you no longer need to do any of that. This is because Google Cloud provides the Vertex AI prediction service, which enables you to host models in production within minutes, using infrastructure that is all managed for you by Google.

For completeness, I will also mention that if you would like to host your models on Google Cloud without using Vertex, numerous other Google Cloud services can be used for that purpose, such as Google Compute Engine (GCE), Google Kubernetes Engine (GKE...