In this paper, the IT-speed dating of a German university is considered, where students have talks with different companies in order to find a suitable internship. The goal is to create a good and fair matching of students and companies for these talks, based on student preferences, and to schedule the resulting talks in order to maintain the given time horizon and minimize the necessary room changes for the students. We solved the problem in two steps. First, we modeled the matching problem as an extended version of the capacitated transportation problem and solved it using a modified stepping stone method. Second, we present two approaches to solve the scheduling problem. A Monte Carlo tree search procedure generates time-constrained schedules with minimal duration, while a genetic algorithm generates longer schedules with individual pauses and fewer room changes. The approaches led to significantly more talks with valuable content, a shorter duration, and greater satisfaction of all participants.