AWS decision-guides documentation change
Summary
Updated content including inference definition, management/deployment comparisons, and expanded feature sections highlighting safeguards and infrastructure controls
Security assessment
Added mention of 'safeguards (guardrails)' in Amazon Bedrock features, which references security controls. However, this appears to be general security feature documentation rather than addressing a specific vulnerability.
Diff
diff --git a/decision-guides/latest/bedrock-or-sagemaker/bedrock-or-sagemaker.md b/decision-guides/latest/bedrock-or-sagemaker/bedrock-or-sagemaker.md index 2b6218177..399b1d22a 100644 --- a//decision-guides/latest/bedrock-or-sagemaker/bedrock-or-sagemaker.md +++ b//decision-guides/latest/bedrock-or-sagemaker/bedrock-or-sagemaker.md @@ -13 +13 @@ IntroductionDifferencesUse -**Last updated** | February 14, 2025 +**Last updated** | June 27, 2025 @@ -23 +23 @@ IntroductionDifferencesUse -Amazon Web Services (AWS) offers a suite of services to help you build machine learning (ML) and generative AI applications. It’s helpful to understand how these services work together to form a generative AI stack, including: +Amazon Web Services (AWS) offers a suite of services to help you build machine learning (ML) and generative AI applications that use [inference](https://docs.aws.amazon.com/bedrock/latest/userguide/inference-how.html), defined as the process of generating an output from an input provided to a foundation model. It's helpful to understand how these services work together to form a generative AI stack, including: @@ -71,0 +72,2 @@ Expertise Required | Basic level of machine learning expertise needed to use pr +Management | Amazon Bedrock provides a simplified API-based approach with minimal infrastructure management. | SageMaker AI requires more infrastructure management, but offers extensive [monitoring](https://docs.aws.amazon.com/sagemaker/latest/dg/model-monitor.html) and [control](https://docs.aws.amazon.com/sagemaker/latest/dg/governance.html) capabilities. +Deployment and Hosting | Amazon Bedrock is serverless, meaning you don't have to manage infrastructure. | SageMaker AI is primarily serverful, and provides granular control over computing resources and scaling. @@ -225,0 +228,19 @@ Amazon Bedrock and Amazon SageMaker AI are optimized for different levels of mac +Features + + +Amazon Bedrock and Amazon SageMaker AI are optimized for different levels of machine learning expertise. + +**Amazon Bedrock** + + * Amazon Bedrock offers a suite of features to help customers build and scale generative AI applications, including model choice features (evaluation), cost and latency optimization features (prompt caching, intelligent prompt routing), customization features (knowledge bases, model distillation), safeguards (guardrails), and agentic features (agents). Amazon Bedrock also offers custom model import, which allows you to import and use customized models with existing FMs through a single, serverless, unified API. + + + + +**Amazon SageMaker AI** + + * With SageMaker AI, you can store and share your data without having to build and manage your own servers. This gives you more time to collaboratively build and develop your ML workflow, and do it sooner. SageMaker AI provides managed ML algorithms to run efficiently against extremely large data in a distributed environment. With built-in support for bring-your-own-algorithms and frameworks, SageMaker AI offers flexible distributed training options that adjust to your specific workflows. Within a few steps, you can deploy a model into a secure and scalable environment from the SageMaker AI console. + + + +