JioTV Content Based Recommendation System
A content based real-time recommendation system serving 300 million views of JioTV
A content based real-time recommendation system serving 300 million views of JioTV
Enhancing the user experience on JioTV, a content based recommendation system was designed to increase the overall user engagement and ease of access and discovery of content across 700+ channels and 20,000 scheduled programming on live "Electronic Programme Guide" served to over 100 million active users every month. The system was designed to be highly scalable with a test load of over 10 billion hits a day and recommendations served at the rate of over 3 million hits per second. This project involved a team of 20 professionals from cross-functional teams like data engineering, devops, user research, tech and analytics.
An interactive result of the recommendation engine after any training event to visualize the functioning of model and to understand how keywords are grouped into topics and relayed to the viewers by understanding their realtime viewing pattern, mood and persona.