ML @turner

By: Josh Kurz

The Journey

We started out with not much knowledge about where we were going and how we were going to get there.

The Tribulations

Once we started working on low level machine learning projects we, found out how far we really had to go and how much we had to learn.

The time we needed to invest into getting ourselves revved up to create production applications was greater than we were comfortable with offering.

The Answer

All of the cloud providers offer high level APIs, which we have decided to use heavily. These API's can help us achieve goals, which we have been thinking about for years. 

We want to strain the API's for everything they offer us. 

Next Steps

We already have four business use cases, from Turner Sports Tech, that we would like to help make a reality.

1. Tagging Media Metadata

2. How to get duplicate clips out of the videos 

3. Automatically caption videos 

4. Quality Control of videos 

Simple POC

We created the Facial Recognition Security application. This is the baseline tech for accomplishing the first use case with NBA.

https://github.com/turnerlabs/rekognition-security