Recommendation Engine for Video Streaming Platform

Project Description

Company Description

We are a young technology company that has pioneered online streaming over mobile devices. We enable our clients to seamlessly stream their content to the consumers though our myplex platform. Our services also include providing advisory to our clients and help them in optimizing their Acquisition and Retention strategy.

Problem Statement

Our vision is to create a best in industry recommendation engine which would aid the content discovery process for users globally by suggesting them content based on their likings and behaviour. The recommendation should be able to use the internally and externally available data about the users/content and leverage sophisticated machine learning models to arrive at relevant recommendations.

The recommendation engine should be able to address capabilities including but not limited to below points:

1. Recommend relevant content to viewers based on their In App behaviour and Demographic details (in case of new user where no other data is available)
2. Create user profile based on their content consumption history
3. Account for recency of the behaviour. Eg: a content viewed in the last 24 hours could be given higher weightage compared to a content viewed 1 month back
4. Provide recommendation based on the time of the day the content is Played/Browsed. Eg:A user might have a habit of consuming news content during the day whereas movies by the evening
5. Provide higher weightage to content viewed for longer duration
6. Capture pure browse behaviour – where the user browses but not consumed
7. For new user the model should be able to learn the preferences of user by asking questions on language, Favourite Genre/Movie/Actor/Director etc
8. Track user preference by the content s/he selects from all the content that is served as recommendation and build it into future recommendations
9. Provide recommendation for different categories. Eg: Recommended for You, Similar Movies, Recommended Movies, Recommended TV Series, New Movies etc
10. To further refine recommendations, along with relevance to user it could also consider overall popularity of the content through internal(usage) as well as external data (imdb, Rotten Tomatoes etc.)
11. Should be able to extract tags for content using external database (Wikipedia, imbd, rotten tomatoes etc.)
12. Our platforms handles millions of content streaming on a monthly basis hence the recommendation engine should be able to process the data and provide real time results
13. Get the user behaviour data in the required format using the APIs provided by the App
14. Should be capable of adding placeholder to manually add content to the carousals in addition to the recommendations when require
15. Provide an interface(Portal as well as App) for users to specify their preferences and showcase recommendation based on the inputs shared by them

Data:

We have multiple clients using the platform and the data available might vary based on information made available by the client. Enclosed are the details of data availability. Our Data is stored in MySQL and Hadoop. We would also like to use other publically available data sources for enrichment. 

Questions that we have:

• We have following Information Request/ questions to gather the required details for considering partnering with your organization to develop the aforementioned Recommendation Engine:
• For which industry were the recommendation engine developed by you earlier? Please share broad description of underlying logic. Also we would like to explore one of the recommendation engines that you have deployed for your clients in the past.
• Profile of experts that will be working on the project along with details of their qualification, past projects and Tools and technologies used earlier?
• What is the technology stack that you would suggest for developing the recommendation engine?
• What approach would you use to develop the recommendation engine?
• How would you suggest to evaluate quality of recommendation?

Timelines

The project needs to be completed within 6 weeks

Please provide a list of milestones and a ballpark estimate of hours to complete each.