Scale AI

Scale accelerates the development of AI applications by helping computer vision teams generate high-quality ground truth data. Its advanced LiDAR, video, and image annotation APIs allow self-driving, drone, and robotics companies to  focus on building differentiated models vs. labelling data.

Companies provide Scale with data via their API and the startup uses its software tools (and team of contractors) to work labelling the text, audio, pictures and video so that its customers’ machine learning models can be trained. The involvement of humans helps lessen labelling bias and provides more accurate information to customers.

Its APIs include: Sensor Fusion, Semantic Segmentation, 2D Boxes/Polygons, Video, 3D Cuboids, Lines & Splines, Keypoints and Data Categorisation. 

In August the startup revealed it had raised a $100m Series C, securing Unicorn status. The round was led by Founders Fund, and had participation from Accel, Index Ventures, Spark Capital, Thrive Capital and more. 

Working with the University of Waterloo in Canada, Scale AI processed a data set for winter driving, to better teach self-driving vehicles how to navigate snow and ice. The data was publicly released in February 2020 to help companies and other universities innovate for such situations.

The startup, led by found Alexandr Wang (who founded the company at 19), has around 100 employees and nearly 30k contractors aiding the labelling process. Its clients include Waymo, OpenAI, Airbnb and Lyft.