he Robot Operating System (ROS) has become the de-facto standard middleware in the robotics community [1]. ROS bundles everything, from low-level drivers to tools that transform between coordinate systems, to state-of-the-art perception and control algorithms. One of ROS's key merits is the rich ecosystem of standardized tools to build and distribute ROS-based software.In a parallel development, a large suite of scientific libraries in the machine learning, deep learning, and data science domains has emerged: NumPy, SciPy, Pandas, OpenCV, the Natural Language Toolkit, PyTorch, and TensorFlow to name a few. These libraries have wide ranging applications including computer vision, computer graphics and natural language processingall of which are also highly relevant in robotics.