Mobile mini–robots are being increasingly used for research and education. Various platforms are available, each focusing on specific key features like CPU or GPU processing, high modularity, or low cost. In this area, the new mini–robot EMAROs (Educational Modular Autonomous Robot Osnabrück) positions itself with a unique combination of modularity, computing capabilities, sensor technologies, and software support. The robot provides two cameras for stereo vision, IMU, distance sensors, and a wide variety of additional sensors and communication interfaces. An important design goal was not only to focus on robotics but also on computer engineering education. Therefore, despite its small size with a diameter of about 10 cm, various computing nodes can be integrated, ranging from embedded CPUs and GPUs to reconfigurable hardware (FPGAs) and dedicated accelerators for machine learning. Using the ROS2 environment, students can focus on algorithm development as well as on resource-efficient implementations, comparing the different computing platforms.