Part 3: Model Management and Security Considerations

In this part, we will understand the core aspects of evaluating and monitoring models, as well as ensuring security and privacy within the Amazon Bedrock environment. Chapter 11 explores techniques for assessing model performance, including automatic model evaluation, human evaluation, and open source tools. Additionally, it covers monitoring techniques such as Amazon CloudWatch, model invocation logging, and integration with AWS CloudTrail and Amazon EventBridge. Chapter 12 centers on data protection, identity and access management, network security, ethical considerations, and implementing guardrails to adhere to responsible AI practices and policies.

This part has the following chapters:

• Chapter 11, Evaluating and Monitoring Models with Amazon Bedrock
• Chapter 12, Ensuring Security and Privacy in Amazon Bedrock