User:GraphicCycle/Matroid

Matroid is a Computer Vision company founded in 2016 in Palo Alto, giving typical computer users the ability to scan video using machine learning.

Product
The Matroid Product is a studio for creating and using detectors, giving users the ability to scan video for faces, actions, places, and other objects. Common examples include finding when former President Barack Obama appears on TV, or how often a man holding a gun is recorded in security footage. Users can easily write a filter—Matroid calls them "detectors"—of their own to find particular people or objects, or they can pick from a library of pre-programmed filters designed by users of the platform.

Matroid is initially focused on users who want to analyze television appearances or scan surveillance video. In the first case, a firm may want to track which political candidates got more TV time or which brand of car appears more often on a particular show or television network. Product features include a video player with computer vision at its core, a tight feedback loop for detector iteration, an extensive API, stream monitoring, and work-sharing amongst users to share detectors.

Scaled Machine Learning Conference
Matroid holds the Scaled Machine Learning Conference (ScaledML) every year, with the previous two conferences held on Stanford University campus. In 2016, Microsoft announced "Open Mind", Microsoft's Visual Studio-like suite for machine learning at ScaledML 2016. At Scaled ML 2017, Google announced the Tensor Processor Unit to speed up machine learning operations in Google Cloud Platform.

Funding
Matroid has raised $3.5 million dollars from New Enterprise Associates in funding in 2016.

Open Source
Matroid uses TensorFlow and Kubernetes, and is an active member of the TensorFlow open source community by holding TensorFlow tutorials and writing a Deep Learning book on the topic with O'Reilly Media. For its contributions to Open Source Machine Learning, at KDD 2016, Matroid was awarded a best paper award.