Summary

In this chapter, we have explored a fully managed cloud solution for serving ML models. You have seen how serving works in Amazon SageMaker, which is a strong representation of a fully managed cloud solution, and you have explored Amazon SageMaker and seen, step by step, how to create a notebook in Amazon SageMaker and how to deploy a model. We have also seen how you can create an endpoint for the model and how you can invoke the endpoint from a client program using boto3. This is our last chapter on the tools that we intended to discuss. There are a lot of tools out on the market and a lot more are coming out. I hope, now that you have an idea about serving patterns, you can choose the right tool for you. Amazon SageMaker is a integral ecosystem for ML engineers and data scientists. This chapter only gives an introduction to serving by building a model from scratch using the models from the model registry. There are many other ways to create models, such as using SageMaker...