As undergraduate enrollment in computer science rises, instructors continue to investigate methods to improve the student experience at scale. One aspect commonly used in courses at scale is queue-driven office hours, in which students join an online queue and meet with teaching assistants on a first-come, first-serve basis (FIFO).

This poster introduces a novel office hours queue feature that automatically groups students in office hours using the cosine similarity metric across their reported issues provided upon joining the queue. Using real office hour attendance data from a 480-person undergraduate course, we find that it is possible to achieve moderate decreases in student wait time during the semester overall (11% on average), with significant decreases possible on the busiest days (20% on average). This approach is suitable for real-world testing and these gains are possible without asking students to provide any additional information than they already do when attending office hours. Therefore, this work provides the basis and motivation for implementation of such an approach in future courses.