Groundlight's platform enables image understanding from natural language queries and a few lines of code.

The company is solving an interesting problem - how to meld AI chatbot interfaces and the large amount of video and imagery that are inside of corporations.  The goal would be for operators to be able to query using text what is happening in video streams, and receive answers to those queries with verification.  Kind of a clever idea.

GeekWire: Amazon and Microsoft vets raise $10M for computer vision startup used in warehouses