Summary
In this chapter, you were introduced to the concept of a CI/CD process as a way to close the gap between building a production-grade ML and getting the model into production. Making use of this methodology, an ML practitioner doesn't simply hand over the trained model to the platform teams but rather integrates the model artifacts into the overall process.
While we haven't as yet shown how the ML practitioner contributes these model artifacts into a process, we have established a pattern of codifying the process by introducing and setting up an AWS CDK project. By using the CDK, we practically demonstrated a backward-working approach for how the engineering team can deploy a trained model as a SageMaker-hosted endpoint CDK construct. We also demonstrated how the engineering teams built the fundamental mechanisms that will eventually automate the integration of the model training and evaluation procedures into the process.
In the next chapter, we will continue...
