Using administrative data from a large, diverse university in California, we identify the extent to which the academic outcomes of undergraduates are affected by the race/ethnicity of their graduate student teaching assistants (TAs) in economics courses. To overcome selection in course taking, we exploit the timing of TA assignments, which occur after students enroll in a course, and use within class and within student variation in TA-student race composition. Focusing on an Asian vs. non-Asian split, results show a positive and significant increase in course grades when students are assigned TAs of a similar race/ethnicity. These effects are largest in classes where TAs were given advanced copies of exams and when exams were not multiple choice. We find positive racial correlations between students and TAs at office hours and discussion sections, suggesting student attendance responds to TA race. We also find some evidence of persistent effects: Racial match improves subsequent student performance in sequenced courses, and positively influences decisions on majoring and future course enrollment for Freshmen and Sophomores. Overall, our evidence is consistent with TA-student match quality gains and role model effects.