pTA: A Programmable Teaching Assistant for Lab Courses

Abstract

Lab courses play a crucial role in enabling students to gain a deeper understanding of theoretical concepts, but these courses require a significant effort from the course’s organizational staff, such as instructors and teaching assistants. To address this challenge, we developed pTA, an acronym for programmable teaching assistant, which automates the functional evaluation of students’ submissions for the Cloud Databases course taught at the Technical University of Munich (TUM). pTA reduces the staff workload and provides instant feedback to students, thereby enhancing their understanding of the project specifications. Additionally, pTA includes a live leaderboard that provides a gamification element that makes the course more interactive and engaging for students. It is deployed on a Kubernetes cluster that ensures scalability with evaluation requests. In this paper, we describe the course’s learning milestones and provide an overview of pTA’s architecture and features. The system’s efficacy was evaluated at TUM and the University of Toronto, where it was deployed in two similar courses. Our findings show that pTA reduced the staff workload by at least 75%, lowered the operating cost, and increased course capacity in terms of the number of students. Furthermore, our study suggests that students exhibit more interest in courses that integrate interactive learning systems and gamification elements.