Machine Learning Researcher
Emphasis: Computer vision, statistical models, neural networks, regressors, optimization methods
Meshcapade is a startup creating realistic human avatars for use in research, apparel, biomechanics, virtual reality and film. Using machine learning and computer vision, we model the nuances of human shape and movement. We build automation from 3D & 4D scans, RGB-D sequences, Mocap, IMU and image data. Our methods derive from SOTA research and we bring realistic human models to life in everyday environments. Our core product, digidoppel, is a consumer-facing platform for the creation, modification, and delivery of our models and related assets.
What we offer
You will work closely with our dev team to help define the technical direction on improving our internal software and methods. You will research, develop and test improvements for our existing research-based code base as well as develop new methods. Our offices are based in Tübingen, Germany. Remote working assistance is also available.
Your position will be at an intersection between 3D computer vision and computer graphics. We are interested in non-rigid shape analysis and statistical modelling of body shapes, and recognition, capture and analysis of motion of humans (and eventually animals too!).
Experience developing in Python and/or C++ (5+ years)
Master's or PhD degree in computer science or another related field
Proven track record of publications on topics related to statistical body modelling, estimation or motion tracking methods
Strong understanding of computer vision and machine learning techniques
Excellent written and oral communication skills
Preliminary experience in related topics such as 3D meshes, point-cloud processing, neural networks, classical optimization, etc.