This thesis aimed to develop a mixed reality-based robotic platform that could support user’s learning, stressing out the differences between traditional learning methods and the learning by teaching paradigm, where the learner reinforces his or her own knowledge by giving lessons to a teachable agent, role in this case played by the Cozmo robot. First, a depth research was conducted to understand the current state of this field of study and the related technology, searching for possible lacks with the aim to take a step forward in the tackled domain. Then, an application was designed and implemented in Unity including a specific mode for each learning paradigm, whose topic was based on the Thèvenin Theorem from the electrical domain, aside from a tutorial to master the user interface.