Contemporary engineering design is characterised by products and systems with increasing complexity coupled with tighter requirements and tolerances. This leads to high epistemic uncertainty due to numerous possible configurations and a high number of design parameters. Set-Based Design is a methodology capable of handling these design problems, by exploring and evaluating as many alternatives as possible, before committing to a specific solution. The Python package PDOPT aims to provide this capability without the high computational cost associated with the factorial-based design of experiments methods. Additionally, PDOPT performs the requirement mapping without explicit rule definition. Instead, it utilizes a probabilistic machine learning model to identify the areas of the design space most promising for user-provided requirements. This yields a plethora of feasible design points, assisting designers in understanding the system behaviour and selecting the desired configurations for further development.