October 5, 2022
We chat with the engineers behind a number of the world’s largest streaming platforms.
SPONSORED BY WARNER BROTHERS DISCOVERY
As we speak’s streaming companies have an unlimited content material library, and a vastly totally different programming problem than conventional tv stations. The place linear TV tries to schedule content material into time slots so the precise demographics will see it and luxuriate in it, streaming companies allow you to watch any content material, any time you need. Meaning these service have to discover a option to ship the precise content material to each consumer, balancing what they may need with what is going to present worth over the long run. That’s the place machine studying is available in.
On this sponsored episode of the podcast, Ben and Ryan discuss with Shrikant Desai and Sowmya Subramanian, two engineering leaders at Warner Bros. Discovery who form how their ML program figures out what your subsequent favourite present is perhaps. We cowl the instruments they’ve been utilizing to construct their studying pipelines, how a viewer’s historical past can form their future, and whether or not ML algorithms can fill a human curator’s function to shock and delight viewers.
Episode Notes
Our friends have accomplished most of their ML work on AWS choices, from AWS Personalize for his or her preliminary advice engine to SageMaker for mannequin coaching and deployment pipeline. Now they’re constructing fashions from scratch in TensorFlow.
Wish to see these suggestions in motion? Take a look at the choices at Discovery+ and HBOMax.
Should you’re a ML/AL knowledge scientist seeking to form the way forward for automated curation, try their open roles.
Observe our friends on LinkedIn:
