Summary
In this chapter, you reviewed the five pillars – operational excellence, security, reliability, performance, and cost optimization – that make up the Well-Architected Framework. You then dove into the best practices for each of these pillars, with an eye to applying these best practices to ML workloads. You learned how to use the SageMaker capabilities with related AWS services to build well-architected ML workloads on AWS.
As you architect your ML applications, you typically must make trade-offs between the pillars depending on your organization's priorities. For example, when getting started with ML, cost-optimization may not be at the top of your mind but establishing operational standards may be important. However, as the number of ML workloads scale, cost-optimization could become an important consideration. By applying the best practices you learned in this chapter, you can architect and implement ML applications that meet your organization&apos...
