Data-Driven Framework for Optimizing Internship Efficiency and Addressing Skills Mismatch in Malaysian Higher Education

The persistent skills mismatch between higher education outcomes and labour market needs presents a significant challenge for graduate employability, particularly in fast-evolving sectors such as artificial intelligence (AI) and data science. This paper introduces a data-driven internship framework designed to bridge this gap by optimising student placements, providing real-time feedback, and ensuring academic curricula remain aligned with labour market demands. The framework integrates three systems: the Student Industry Placement System (SIPS), which uses machine learning to match students with relevant internships; the Intern Learning Management System (ILMS), which tracks skill development through continuous feedback loops; and the Industry Profiling System (IPS), which leverages predictive analytics to provide real-time labour market insights. The framework was evaluated through a mixed-methods approach, incorporating pilot program data from five universities and 15 industry partners and simulated data based on historical internship trends. The SIPS algorithm achieved a placement accuracy of 89%, significantly improving over traditional methods (60-65%). 78% of students received continuous feedback through the ILMS, resulting in measurable improvements in technical skills (30% increase in coding proficiency) and problem-solving abilities (25% improvement). The IPS enabled universities to update 15% of their curricula in response to real-time labour market data, particularly in high-growth sectors such as AI and digital transformation. 68% of the students secured full-time employment within six months of graduation, with 65% in roles directly aligned with their internships. These findings demonstrate the framework's scalability and potential for reducing skills mismatch, making it a viable model for other emerging economies.