Computer Vision Engineer - Body Sensor Networks

Do you get excited about building cutting-edge perception systems that work reliably in the real world? Do you have a strong foundation in computer vision, robotics, or sensor fusion? Are you interested in shaping new technology capabilities within a growing and innovative high-tech company, where people and fun are part of our core values? We’d love to have you! Who is Xsens Xsens is a leading full-stack provider of sensors, software, and analytics that enable the digitization of movement. Xsens serves entertainment, humanoid robotics, health & sports and automation & mobility markets. By turning complex motion data into actionable insight, we help our partners accelerate development, improve control and learning, and bring physically intelligent systems to life. We collaborate with leading global innovators including Boston Dynamics, Toyota, Siemens, Electronic Arts, Epic Games, Netflix, and 20th Century Studios, alongside more than 1,000 research institutions, humanoid robotics companies and universities worldwide. Your role as a Computer Vision Engineer at Xsens As a Computer Vision Engineer, you will lead camera-based perception for Xsens motion capture. Your role is to establish Deep Learning and Computer Vision pipelines to turn camera images into human pose estimate. You will learn from the highly-skilled team of sensor fusion experts, biomechanists and mathematicians. You will be its computer vision specialist, and report to the Technology Manager. The position is open to hybrid / remote working. Key challenges: - Select, adapt and fine-tune deep learning models for 2D and 3D human pose estimation from cameras, that perform robustly in challenging real-world conditions. - Develop camera models and calibration routines for a range of camera and lens configurations. - Evaluation of motion capture output and developing compensatory algorithms and techniques to handle camera /vision failure modes. - Build training and evaluation pipelines from camera and optical motion capture recordings. Use synthetic data where real data is insufficient. - Track the computer vision research landscape and assess whether new techniques are feasible for Xsens products.