Context. This thesis creates a learning unit for teaching decision trees to PhD students whose degrees were not computer science oriented. Problem. Due to the fact that Machine Learning has great benefits for researching topics, more and more PhD students outside the field of informatics want to use it for their own research and therefore need a basic understanding of the field to get started. Objective. The objective of this thesis is to create the outlines of a learning unit for PhD students without knowledge in Machine Learning based on pedagogic and didactic research. The learning unit will use decision trees as an example. Furthermore it builds a base for teaching other Machine Learning content. Method. This thesis surveys research done on teaching STEM content and compares and adapts it to create a fitting learning unit for this thesis’s objective. Result. This thesis provides an approach to teaching the basics of decision trees as an example of Machine Learning content. It uses the Berlin Model and an Active Learning design. The core design of this learning unit represents a first concept to teaching further Machine Learning content to PhD students without an appropriate background.