SAMUR.A.I: A Curriculum-Aligned Retrieval-Augmented Tutoring Agent

Rayleen Ramos

Co-Presenters: Luis Velazquez Rodriguez, Maria Santos, Christian Zambrano

College: Hennings College of Science Mathematics and Technology

Major: MS.COMPUTER/SCIENCE

Faculty Research Mentor: Kwak, Daehan  

Abstract:

Limited faculty office hours, constrained tutoring availability, and the growing demand for flexible academic support create significant barriers for students seeking timely help. While general-purpose AI tools exist, they lack alignment with specific university curricula and risk providing inaccurate or misleading information. To address these challenges, this project presents SAMUR.A.I., a curriculum-aligned AI tutoring and self-study chatbot designed specifically for Kean University students. SAMUR.A.I. leverages Retrieval-Augmented Generation (RAG) combined with Large Language Models (LLMs) to deliver accurate, course-specific assistance.The system is trained exclusively on Kean-approved academic materials, including course syllabi, lecture slides, assignments, rubrics, and instructional policies from computer science courses such as CPS 1231 and CPS 2231. When a student submits a question, the system first retrieves relevant curriculum content. Then it generates a response grounded in that material, reducing hallucinations and ensuring alignment with how courses are taught at Kean. In addition to conceptual academic support, SAMUR.A.I. incorporates guided coding assistance tailored to programming coursework. Rather than immediately generating full solutions, the system employs a scaffolded, hint-based approach that promotes active learning. Students receive step-by-step explanations, debugging guidance, and incremental problem-solving prompts designed to mirror instructional best practices. This ensures learners build foundational understanding and computational thinking skills while maintaining academic integrity. By balancing support with pedagogical responsibility, the platform functions not only as an answer engine but as a developmental learning companion.The chatbot is designed to provide 24/7 academic support, enabling students to learn at their own pace and engage with course material outside of traditional classroom and office hour constraints. By offering instant, curriculum-specific explanations, SAMUR.A.I. aims to improve student confidence, understanding, and academic performance. For faculty and supplemental instruction leaders, the system helps reduce repetitive clarification questions while maintaining consistent, institution-approved guidance. Overall, SAMUR.A.I. demonstrates how responsible, domain-specific AI can enhance accessibility, support student success, and promote effective learning in higher education.