This paper exploits random assignment of students to university instructors to investigate the effect of instructor gender on student performance, course preference and labor market outcomes. I find that female instructors increase females’ performance and probability to take follow-up courses in the same subject. In the longer-run instructor gender does not seem to contribute to explaining occupational sorting in the labor market. However, women, who experienced more female instructors in their first university year are less likely to be unemployed after entering the labor market.