OptiTeam: AI-Augmented Team Formation for Capstone Courses¶

Conference: TPN (Teacher-Participant Network) 2026
Type: Poster Presentation
Authors: Michael Pisman, Santosh Chandrasekhar
Links: GitHub

Abstract¶

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. OptiTeam introduces a data-driven methodology that combines Large Language Models (LLMs) with Integer Linear Programming (ILP) to form optimal student teams. The system first uses LLMs to extract skill requirements from project descriptions, then collects student preferences and self-assessed skills via survey, and finally applies ILP to simultaneously maximize skill coverage and preference alignment across all teams. The poster presents the system architecture, the optimization formulation, and preliminary results from deployment in CSE120 capstone courses at UC Merced.