Recommendation Engine in Live Streaming: how to recommend the unknown

THE PANEL DISCUSSION IS NOT RECORDED

Every day in Tango, more than 50 thousand unique broadcasters make more than 250 thousand stream sessions which generates over 10 petabytes of data per month. Daniel will talk about how to build an effective runtime system of broadcast recommendations from scratch, constantly decreasing the response time to user feedback. What algorithms are used for this, and he will be happy to answer all the listener's questions.