Machine Learning Approach for Predicting the Career Paths of IT Undergraduates

Abstract

Career trajectories of Information Technology (IT) undergraduates are crucial for aligning academic programs with industry demands and supporting students' professional aspirations. Rapid changes in the IT industry present both opportunities and challenges for new graduates. Despite the availability of information, many IT undergraduates struggle to secure jobs that match their aspirations and abilities. Traditional career development frameworks often fail to provide accurate predictions due to their inability to address industry-specific requirements, technical skills, and soft skill demands, leading to graduate unemployment and skills imbalances. To address these challenges, this research proposes a machine learning (ML)-based framework for predicting suitable career paths for Sri Lankan IT undergraduates by analysing a combination of these features to tailor career predictions to individual profiles. The study focuses on six career fields: Software Engineering, UI/UX Design, Quality Assurance, Business Analysis/Project Management, Data Science/Artificial Intelligence, and Networking/DevOps/System Administration. Through an extensive literature survey and expert opinions, 49 attributes were identified, including technical skills, soft skills, academic performance, and internship experiences. A dataset of 520 IT professionals was collected, pre-processed, and used to train and evaluate eight ML models: XGBoost, Artificial Neural Network, Decision Tree, Random Forest, SVM, Gradient Boosting, AdaBoost, and an Ensemble model. The Ensemble model, combining Random Forest, Gradient Boosting, and SVC in a stacking approach, achieved the best performance with 91.39% accuracy, 92.47% precision, 91.39% recall, and a 91.48% F1-score. This framework seamlessly integrates into career development systems, such as university counselling platforms and job portals, by enhancing traditional methods with data-driven insights. It highlights the impact of key features, ensuring accessibility for educators, policymakers, and students. By bridging the gap between academia and industry, this research empowers IT undergraduates to make informed career decisions, enhancing job satisfaction and industry alignment.