Master Thesis in Reinforcement Learning for Multiple-Embodiment Grasping

During your Master thesis, you will design architectures and pipelines to transfer current methods to an online RL algorithm. You will create RL environments for multi-embodiment grasping using our grasp dataset generation pipeline. Furthermore, you will benchmark the current state of the art approaches in robotics grasping. Finally, you will work under the remote and on-site supervision of the research staff, who will guide you throughout your thesis. Most of the systems and frameworks are already in place, giving you an excellent opportunity to hone your skills.