Which AWS service allows for developing, testing, and deploying machine learning models without managing infrastructure?

Amazon SageMaker is the service designed to simplify the process of developing, testing, and deploying machine learning models while taking care of infrastructure management. It provides a fully managed environment where users can build, train, and deploy machine learning models at scale.

One of the key strengths of Amazon SageMaker is that it abstracts much of the underlying infrastructure and complexities often associated with machine learning workflows. Users can leverage built-in algorithms and frameworks as well as easily integrate their own custom models without needing to worry about provisioning servers or managing scalability issues. The service also offers features for data labeling, algorithms selection, training management, and model evaluation, all within a unified interface.

In contrast, the other services mentioned focus on specific functions or areas within the AWS ecosystem. For example, Amazon Comprehend specializes in natural language processing tasks, Amazon Rekognition is aimed at image and video analysis, and Amazon Lambda facilitates serverless computing for running code in response to events. While these services provide valuable capabilities, they do not offer the comprehensive environment for creating machine learning models that Amazon SageMaker does.