Amazon Personalize

2019: Emergence in a public access

In June, 2019 released Amazon in a public access the Personalize service allowing to train and implement models of machine learning for formation of personal recommendations. 

Amazon Personalize represents completely managed service which trains, configures and implements the user models of machine learning in a cloud, providing necessary infrastructure and managing such things as data processing, extraction of functions, a hosting, training and optimization of algorithms.

Amazon started service for embedding of "smart" recommendations in the websites and mobile applications

Clients provide to service activity flows (the number of clicks, viewings pages, authorizations, purchases) in the applications and the websites and also the list of things which they want to recommend (articles, products, videos, audio recordings, etc.). Besides, for drawing up more exact recommendations it is possible to specify age of target audience, the geographic location and other additional parameters.

The company receives results of information processing through API. At the same time payment is made only for what was used by the client.

Through Amazon Personalize it is possible to embed AI models in  mobile applications, the websites, content management systems and  email marketing, It is possible to build in algorithms for personalisation of recommendations of products, search results and  direct marketing.

As the vice president for machine learning in division of Amazon Web Services (AWS) Swami Sivasubramanian reported, "the best part" of Amazon Personalize was the fact that developers do not need experience in area of training for a training, settings and deployments of "smart" algorithms according to business challenges.

Affirms that Amazon Personalize lowers by 60% installation time and  settings of infrastructure and  algorithms for models of machine learning.^[1]

Notes