Samenvatting

The aim of this thesis is to identify and implement the best personalized
recommendation system that will increase the engagement of Fitenium users. Based on
the research, it was determined that the best personalized recommendation system for
this use case is Amazon Personalize.
Amazon Personalize makes it easy for developers to create personalized
recommendation systems. The recommendation system was designed to be integrated
in the existing architecture of the Fitenium project. With the help of Lambda functions
and Step Functions, an automatic workflow of retraining and deploying the machine
learning model was designed.
The design was implemented by developing all the individual Lambda functions
and creating the state machine that represents the workflow for retraining the model.
The infrastructure was created using the Infrastructure as Code paradigm, with the help
of AWS CloudFormation and Serverless Application Model.
The solution was deployed and tested in production mode. Analytical data
showed that the new personalized recommendation algorithm improved the conversion
of the following recommendations to 4.7%, which is almost 10 times more compared to
the previous heuristic solution.