Buzzwords: CNN, ML, collaborative filtering, multi-armed bandits
There are currently almost 15,000 VR videos in SLR Premium. 80% of SLR's views are coming from recommended videos. Better algorithms help our user to find videos they like faster, in the app and on the web, it also lets them discover new videos that they could have missed.
We are looking for a recommender system specialist to join our team and help us elevate our suggestions to the next level, what we have now is a mix a collaborative filtering and content-base recommendations, any changes to current design are run thru classic AB testing.
If you can take it a step further, please email me at philip @ sexlikereal.com with "Video recommendations" title
Some ideas where things might be going to from here for us:
Testing faster for a faster rollout?
https://www.youtube.com/watch?v=kY-BCNHd_dM&ab_channel=DataCouncil
Looking inside the video for spatial data?
https://benanne.github.io/2014/08/05/spotify-cnns.html
The Netflix Recommender System: Algorithms, Business Value,
and Innovation
https://dl.acm.org/doi/pdf/10.1145/2843948
