Summary
This chapter introduced AWS and Amazon SageMaker as a platform for building and deploying ML solutions. An overview of the SageMaker service was given, including the Clarify service, which provides advanced features such as model bias checks and explainability.
We then proceeded to build a complete ML pipeline with the SageMaker service. The pipeline includes all steps of the ML life cycle, including data preparation, model training, tuning, model evaluation, bias checks, explainability reports, validation against test data, and deployment to cloud-native, scalable infrastructure.
Specific examples were given to build each step within the pipeline, emphasizing full automation, looking to enable straightforward retraining and constant monitoring of data and model processes.
The next chapter looks at another MLOps platform called PostgresML. PostgresML offers ML capabilities on top of a staple of the server landscape: the Postgres database.
