Brandon Pardi, Garret Castro, Michael Pisman, Avash Adhikari, Santosh Chandrasekhar (2026) — CSCI 2025 (12th International Conference on Computational Science and Computational Intelligence) 📄 Full text (arXiv)
Team-based projects are a cornerstone of engineering and computing courses, but unstructured team formation often leads to poor project outcomes due to misaligned student interests and inadequate skill coverage. This paper introduces a novel, three-stage methodology for creating effective student teams by integrating student preferences with project skill requirements. Students complete a survey, an LLM analyzes project descriptions to extract skills, and a dynamic assignment algorithm matches students to projects. Preliminary evaluations show higher skill coverage and preference satisfaction compared to random or manual assignment, overcoming limitations of CATME Team-Maker.
Key Contributions
- LLM-driven team formation outperforms CATME Team-Maker for skill coverage and preference alignment in capstone courses.
Related Pages
- ai-k12-evidence-base — Empirical evidence on AI in K-12 education outcomes
- intelligent-tutoring — Automated tutoring systems and their evaluation
- student-experience — Student perspectives on AI in education
- learning-analytics — Data-driven approaches to understanding learning
- llm-feedback-programming-classroom — LLM feedback in classroom settings
Citation
APA: Brandon Pardi, Garret Castro, Michael Pisman, Avash Adhikari, Santosh Chandrasekhar (2026). Improving Capstone Team Outcomes through Dynamic Skill Matching and Preference Alignment. arXiv:2606.15572. CSCI 2025 (12th International Conference on Computational Science and Computational Intelligence).