The second service layer is AWS Sagemaker, which is designed to handle the entire workflow. This is more powerful than prebuilt services with the option to develop custom algorithms or to leverage existing training sets and AWS algorithms.

Its modular architecture makes Sagemaker flexible to any use case, allowing you to utilise SageMaker independently for model building, training or deployment and to accelerate the development of the business solution.

Leveraging AWS can reduce your total cost of ownership and time to deployment. This is most obvious in smaller applications where Sagemaker can reduce the TCO by up to 90% with no loss of functionality or flexibility by using a fully managed service.

The third option is the AWS Deep Learning AMIs which provide the infrastructure and tools to accelerate deep learning in the cloud, at any scale, enabling complete customisation.

Using Deep Learning AMIs we can quickly launch Amazon EC2 instances pre-installed with popular deep learning frameworks and interfaces such as TensorFlow, PyTorch, Apache MXNet, Chainer, Gluon, Horovod, and Keras to train sophisticated, custom AI models. Whether your use case will benefit from Amazon EC2 GPU or CPU instances, there is no additional charge for the Deep Learning AMIs – you only pay for the AWS resources needed to store and run your applications.