Amazon has announced API access to the AI that drives much of Amazon.com through Amazon Personalize. The machine learning service allows contextual personalization and recommendations for third parties. Retailers are the most obvious target since Amazon has spent over 20 years perfecting its retail-focused AI, but any site or app that offers user recommendations through search results, email campaigns, or otherwise can benefit from Personalize.
Developers do not need much machine learning experience to utilize the service. Developers can build and train the service through the AWS console or using the AWS SDK. Developers need only to format input data and upload to an Amazon S3 Bucket, select a training algorithm, train the solution and deploy the solution. Alternatively, Personalize can capture live data from real users and make recommendations accordingly.
Personalize includes three related components. Amazon Personalize is used to create, manage, and deploy solutions. Amazon Personalize Events is used to record user events for further training. Amazon Personalize Runtime is used to receive recommendations from a deployed solution. To learn more, visit the prerelease docs.