Create a Free Account

Amazon AI services bring natural language understanding (NLU), automatic speech recognition (ASR), visual search and image recognition, text-to-speech (TTS), and machine learning (ML) technologies within the reach of every developer. Based on the same proven, highly scalable products and services built by the thousands of deep learning and machine learning experts across Amazon, Amazon AI services provide high-quality, high-accuracy AI capabilities that are scalable and cost-effective.

In addition, the AWS Deep Learning AMI provides a way for AI developers and researchers to quickly and easily begin using any of the major deep learning frameworks to train sophisticated, custom AI models; experiment with new algorithms; and learn new deep learning skills and techniques on AWS’ massive compute infrastructure.

Amazon AI Product Strategy

Our approach to AI is made up of three main layers that sit on top of the AWS infrastructure:

AI Services: At the highest level, for developers who want access to AI technologies without having to train or develop their own ML models, AWS provides a collection of highly scalable pre-trained and pre-tuned managed AI Services that do not require any previous artificial intelligence or deep learning knowledge in order to get started. Amazon Rekognition for image and facial analysis, Amazon Polly for text-to-speech, and Amazon Lex for building conversational chatbots with automatic speech recognition and natural language understanding (NLU) capabilities.

AI Platforms: For customers with existing data who want to focus on building custom inference models, we provide a set of AI platforms which remove the undifferentiated heavy lifting associated with deploying and managing AI training and model hosting. The Amazon Machine Learning service allows you to train custom machine learning models using your own data, without requiring deep machine learning skills or expertise. In addition, Apache Spark on Amazon EMR includes MLlib for scalable machine learning algorithms.

AI Frameworks: Finally, we support all major AI frameworks for researchers and data scientists who want to build sophisticated and cutting-edge intelligent systems. Frameworks such as Apache MXNet, TensorFlow, Caffe, Theano, Torch, Keras, and CNTK provide flexible programming models for training custom models at scale. The AWS Deep Learning AMI, available for both Amazon Linux and Ubuntu, provides all of these frameworks pre-installed and configured on a convenient Amazon Machine Image to help you get started quickly and easily.

AI Infrastructure: Deep learning frameworks, like Apache MXNet, use neural nets, which involve the process of multiplying a lot of matrices. Amazon EC2 P2 instances provide powerful Nvidia GPUs to substantially accelerate the time to complete these computations, so you can train your models in a fraction of the time required by traditional CPUs. After training, Amazon EC2 C4 compute-optimized and M4 general purpose instances are ideally suited for running inferences with the trained model. In addition, AWS Lambda lets you simplify your operations with serverless machine learning predictions, while AWS Greengrass lets you run AI IoT applications seamlessly across the AWS Cloud and local devices.

Back to Top »

Amazon AI services make it easy for any developer to build applications that can turn text into lifelike speech; have conversations using voice or text; and analyze images to identify faces, objects, and scenes.

Amazon Lex

Amazon Lex uses the same technology as Amazon Alexa to provide advanced deep learning functionalities of automatic speech recognition (ASR) and natural language understanding (NLU) to enable you to build applications with conversational interfaces, commonly called chatbots.

Learn more »

Amazon Polly

Amazon Polly is a service that turns text into lifelike speech. Polly lets you create applications that speak in over two dozen languages with a wide variety of natural sounding male and female voices to enable you to build entirely new categories of speech-enabled products.

Learn more »

Amazon Rekognition

Amazon Rekognition, built on technology used by Amazon Prime Photos to analyze billions of images daily, is a service that makes it easy to add image analysis to your applications. With Rekognition, you can detect objects, scenes, and faces in images, as well as search and compare faces between images.

Learn more »

For developers with existing data who want to build custom models, the AI Platform services remove the undifferentiated overhead associated with deploying and managing AI training and model hosting.

Amazon Machine Learning

Amazon Machine Learning provides visualization tools and wizards that guide you through the process of creating machine learning (ML) models without having to learn complex ML algorithms and technology.

Learn more »


Apache Spark on Amazon EMR includes MLlib to deploy scalable machine learning algorithms, or you can use your own libraries. By storing data sets in-memory, Spark can provide great performance for machine learning applications.

Learn More »

The AWS Deep Learning AMI (available for Amazon Linux and Ubuntu) and the AWS Deep Learning CloudFormation Template let you quickly deploy and run any of the major deep learning frameworks at any scale. The AWS Deep Learning AMI lets you create managed, auto-scaling clusters of GPUs for large scale training, and run inference on trained models. It is pre-installed with Apache MXNet, TensorFlow, Caffe2 (and Caffe), Theano, Torch, CNTK, and Keras. The AWS Deep Learning AMI is provided and supported by Amazon Web Services, for use on Amazon EC2. There is no additional charge for the AWS Deep Learning AMI – you only pay for the AWS resources needed to store and run your applications.

Apache MXNet is Amazon’s deep learning framework of choice and is the platform for our AI services, as well as many AI projects within It is a flexible, efficient, portable, and scalable open source library for deep learning that supports declarative and imperative programming models across a wide variety of programming languages and use cases.

Apache MXNet features a single implementation of backend system and common operators with support for a large number of frontend languages, including Python, C++, Scala, and R. Due to Apache MXNet’s architecture, performance remains consistent regardless of the frontend language used.

Unique memory optimizations allow Apache MXNet to be used across a wide variety of use cases. After taking advantage of the cloud to train your model, they can be deployed in connected devices at the edge, mobile phones, browsers, industrial and consumer drones, or simply remain in the cloud.

Apache MXNet inherently supports automatic scheduling of portions of source code that can be parallelized over a distributed environment. Paired with Amazon EC2 P2 instances, Apache MXNet applications scale across GPUs with up to 91% efficiency, and across cluster nodes with up to 88% efficiency. 

Apache MXNet

TensorFlow is an open source software library for numerical computation using stateful dataflow graphs.

Caffe2 is a lightweight, modular, and scalable deep learning framework designed to help researchers train large machine learning models and deliver AI on mobile devices.

Keras is a high-level neural networks library, written in Python and capable of running on top of either TensorFlow or Theano. It was developed with a focus on enabling fast experimentation. 

The Microsoft Cognitive Toolkit is a unified deep-learning toolkit by Microsoft Research that describes neural networks as a series of computational steps via a directed graph.

Torch is a scientific computing framework with wide support for machine learning algorithms that puts GPUs first. It is easy to use and efficient, thanks to an easy and fast scripting language, LuaJIT, and an underlying C/CUDA implementation. 

Theano is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. 

Amazon Machine Images are a great way to quickly start using deep learning technologies on AWS. The AWS Deep Learning AMIs come pre-installed with popular open source deep learning frameworks (Apache MXNet, TensorFlow, Theano, Torch, CNTK and Caffe), GPU-acceleration through pre-configured CUDA drivers, and supporting tools such as Anaconda and Jupyter.

To learn more, visit the AWS Deep Learning AMI site. 

Cloud Formation Template

AWS CloudFormation templates are an easy way to scale up multiple instances of EC2 instances for big compute jobs such as training deep neural networks. Developers can use the distributed Deep Learning CloudFormation template to spin up a scaled-out, elastic cluster of P2 or G2 instances using the Deep Learning AMI for their larger training requirements.

To learn more, visit the AWS EC2 Compute Blog for information on CloudFormation usage for Deep Learning.