One of the Leading Healthcare Product Company intended to improve patient satisfaction by recommending videos based on various parameters

The Challenge

Legacy recommendation engine driven based on pre-defined set of rules, that recommends users based on preset interest category or Health service provider’s preset video options.

Client needed to move towards a much more personalized solution that takes into account Realtime feedback and makes recommendations accordingly



Our Solution

The solution is deployed across various Health care Institutions where the videos are aggregated from content providers and stream them for the end user;

Evaluated AWS Personalize Vs. Custom Solution and Created connectors with client data sources by defining glue data catalog

Raw Data is dumped to s3 and further processed to publish to Dynamo DB and Models were trained on CPU based EC2 Compute instances and served based on docker pod scaling architecture

An API that generates recommendation scores for the videos watched by the

A feedback loop that aggregated changes to
user/video meta data and re-calibrates itself as a batch process



  • Results

  • 42%

    reduction in exit / bounce rates


  • 40%⬆

    minutes Spent on the App


  • 33%⬆

    length of the videos watched