Simon Walser, Talia Arauzo, Brian Fu, Sailesh Mohapatra, Sam Deverett, Jeremy Berkowitz

Berkeley, California, Dec 11, 2020 – uses computer vision to accurately identify body movements and classify exercises performed by patients in real-time. The company announced the launch of its product, designed for physical rehabilitation clinics in Europe and North America, early this November, and is already making waves in the physiotherapy community., which traces its origins to UC Berkeley’s Sutardja Center for Entrepreneurship and Technology DataX course, was spun out of a semester-long project in collaboration with Guido Schuster’s Interdisciplinary Center for Artificial Intelligence.

Physical rehabilitation is a growing pain in an aging population. At the same time, therapy tends to be slow, expensive, and constrained by the availability of trained rehabilitation doctors and specialists.’s computer vision-based approach utilizes statistical models based on both static and dynamic analyses to classify one’s exercise movements as correct or incorrect. This real-time movement evaluation reduces reliance on doctors to plan and monitor every aspect of therapy and recovery, while generating a rich source of data to map progress made by the patient. This translates into more personalized care, faster recovery, and more patient support. At the same time,’s physical rehabilitation clinics will be able to process more patients daily and track every aspect of their therapy with more data than ever before. This approach could be a gamechanger for the accessibility and safety of physical therapy.

“Kinesis makes technology that was previously only available to top professional athletes accessible to everyday patients,” says CEO Sam Deverett. “At the same time, it creates an incredibly powerful new way to monitor rehabilitation without the traditional roadblocks of staff, scheduling, or compromises to patient experience.” is easy to install and calibrate, as well as is infinitely scalable to any exercise that can be performed within a cubic eight-foot domain. Once installed, it learns any exercise demonstrated by therapists and highlights pose, speed, or body part misalignments performed by the patient. The patient then leaves with a calculable set of adjustments to their physiotherapy exercises, which they can implement at home before their next appointment, significantly advancing their progress.

About Us is an early-stage start-up spun out of UC Berkeley in association with OST Switzerland. The team of five is advised by researchers from OST and UC Berkeley. Their vision is to use advances in AI/ML to revolutionize the way physical rehabilitation is administered worldwide.