Specialist interviews with specific hints play an essential role across various fields, including education. These evaluations offer to evaluators a clear insight into the depth of understanding achieved in a particular subject. Subjectivity and inaccurate tracking can hinder effective assessment, especially in fields like psychology, social sciences, or literature. Thus, in mathematical sciences or engineering, where computations and formulas are involved, identifying mistakes can also be challenging. The purpose of this work is to develop an application (LEEM - Learning Error Evolution Matrix) as a modern and analytical instrument for analyzing and classifying errors in these kinds of interviews. By identifying the conceptual roots and chronological nature of mistakes that are made by candidates, LEEM app aims to distinguish between temporary misunderstandings and long-term knowledge gaps. The goal is to provide reviewers with a diagnostic framework that enhances evaluation accuracy and guides more effective training. The mistakes are mapped in an error evolution matrix similarly to the concept behind the Eisenhower Matrix.