People count through a door might sound simple. Solved, even. In reality, there are significant and unexpected challenges. Our sensor’s analysis must account for millions of environmental variates; handle odd and unpredictable human behavior; and, to be ubiquitous, must amass its data anonymously. The problem of people count is unsolved and complex; but we’re convinced it’s fundamental, fascinating, and global. At the core of our method is an image processing algorithm. It understands the direction of objects as they collide and move through a threshold, it’s capable of discerning a door swing from a plant, and it knows when a line has formed. We’re looking for an experienced computer vision engineer to help us turn large amounts of data into real-time, always-on people count. Ideally, you have a deep interest in computer vision, especially in segmentation and tracking, and experience architecting vision-based products. If you get excited about predictive movement analysis, background segmentation, and optimizing code to run on an embedded device, we’d love to talk.