An AI-Powered Framework for Skill Assessment and Personalized Career Pathway Recommendation in STEM and Coding Courses for High School Students: A Systematic Literature Review
DOI:
https://doi.org/10.64252/vambja45Keywords:
Artificial Intelligence in Education, STEM Skill Assessment, Career Pathway Recommendation, Hybrid Recommender Systems, Explainable AI in SchoolsAbstract
Background: Artificial Intelligence (AI) is increasingly applied in secondary Science, Technology, Engineering, and Mathematics (STEM) and coding education to enhance skill assessment and guide career decision-making. By analysing learner data, AI frameworks can identify strengths, address skill gaps, and recommend personalised academic and vocational pathways. However, integration into unified, pedagogically aligned systems remains limited, and research activity is unevenly distributed across regions.
Objective: This review synthesises peer-reviewed literature on AI-powered frameworks that combine skill assessment with career pathway recommendation in secondary-level STEM and coding education, with a focus on methodological trends, geographic distribution, and pedagogical implications.
Methodology: A systematic search, following PRISMA 2020 guidelines, was conducted across Scopus, Web of Science, IEEE Xplore, ACM Digital Library, and ERIC for publications from 2010 to 2025. Studies were eligible if they were peer-reviewed, published in English, and applied AI techniques, such as machine learning, deep learning, or other algorithmic methods, for skill assessment or career recommendation in secondary STEM contexts.
Results: Machine Learning was the dominant AI technique (50.0%), while hybrid recommender systems integrating content-based and collaborative filtering were the most prevalent career guidance approach (40.0%). Research output was concentrated in the Middle East (25.0%) and Asia-Pacific (25.0%), followed by Europe (15.0%), with Latin America and North America each contributing 10.0% and Africa 5.0%. Explainable AI was present in 10.0% of studies, indicating gradual but limited adoption.
Conclusion: AI demonstrates strong potential to enhance differentiated instruction, early skill-gap detection, and personalised career guidance in secondary STEM education. To maximise impact, future research should address geographic imbalances, embed pedagogical theory into system design, and advance interpretable multi-modal hybrid models adaptable to diverse educational contexts.
