A way to keep your models up to date without re-training entirely new models every time something major happens in the world.

A small team out of San Francisco trying to implement 'Real-Time Machine Learning'