TY - GEN
T1 - LLM-Assisted Software Development in Aerospace Engineering Education
T2 - 2026 International Conference on Artificial Intelligence, Systems, and Emerging Technologies, ICAISET 2026
AU - Kafafy, Raed
AU - Okasha, Mohamed
AU - Abdelkhalik, Ossama
N1 - Publisher Copyright:
© 2026 IEEE.
PY - 2026
Y1 - 2026
N2 - Large language models (LLMs) are increasingly being applied to software development tasks; however, their structured integration into aerospace engineering workflows remains insufficiently documented. This paper presents a structured workflow for LLM-assisted software development in aerospace engineering education, emphasizing problem decomposition, modular prompt design, iterative validation, and human-in-the-loop verification.The proposed methodology is demonstrated through a case study involving the development of a web-based visualization tool for orbital mechanics instruction. Rather than focusing on the software architecture itself, the study analyzes the interaction between domain expertise and AI-generated code, identifies common error categories, and evaluates the role of structured prompting in improving reliability and development efficiency. The results highlight both the potential and limitations of LLM-assisted engineering, underscoring the necessity of expert validation in physics-based applications. The presented workflow provides a practical framework for integrating large language models into aerospace software development and educational tool prototyping.
AB - Large language models (LLMs) are increasingly being applied to software development tasks; however, their structured integration into aerospace engineering workflows remains insufficiently documented. This paper presents a structured workflow for LLM-assisted software development in aerospace engineering education, emphasizing problem decomposition, modular prompt design, iterative validation, and human-in-the-loop verification.The proposed methodology is demonstrated through a case study involving the development of a web-based visualization tool for orbital mechanics instruction. Rather than focusing on the software architecture itself, the study analyzes the interaction between domain expertise and AI-generated code, identifies common error categories, and evaluates the role of structured prompting in improving reliability and development efficiency. The results highlight both the potential and limitations of LLM-assisted engineering, underscoring the necessity of expert validation in physics-based applications. The presented workflow provides a practical framework for integrating large language models into aerospace software development and educational tool prototyping.
KW - aerospace engineering education
KW - AI-assisted software development
KW - human-AI collaboration
KW - large language models (LLMs)
KW - prompt engineering
KW - structured workflow
UR - https://www.scopus.com/pages/publications/105042179859
UR - https://www.scopus.com/pages/publications/105042179859#tab=citedBy
U2 - 10.1109/ICAISET66439.2026.11541825
DO - 10.1109/ICAISET66439.2026.11541825
M3 - Conference contribution
AN - SCOPUS:105042179859
T3 - 2026 International Conference on Artificial Intelligence, Systems, and Emerging Technologies, ICAISET 2026
BT - 2026 International Conference on Artificial Intelligence, Systems, and Emerging Technologies, ICAISET 2026
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 21 April 2026 through 23 April 2026
ER -