SkillBridge: A Job-Course Matching Framework Using Pre-Trained Language Models and Graph Representation

In today's rapidly evolving job market, it is crucial for universities to customize the curriculum of their programs to meet the ever-changing needs of the industry. However, the alignment of job candidates with suitable career paths remains a complex challenge, primarily due to the heterogeneous nature of job descriptions and applicant profiles. By aligning educational offerings with market demands, colleges and universities can design and adapt their curricula to meet the specific demands of the job market, ensuring graduates are equipped with the right competencies. Individuals, in turn, can receive tailored career recommendations based on the skills they have gained through their academic programs, helping them identify careers that align with their knowledge and skills. Building on the need for aligning educational offerings with market demands, this research proposes a novel approach, named SkillBridge, to entity matching in heterogeneous datasets. Utilizing pre-trained language models, the method addresses the challenge of unlabeled datasets, which are common in real-world scenarios due to the high cost and time required for labeling. Additionally, graph representation is employed to provide visual explanations and insights into the relationships between entities, enhancing the interpretability of the proposed model. The effectiveness of this approach is demonstrated through its application to job market vacancies and university courses, where it achieves a significant improvement of about 20 % in the F1 score metric compared to the baseline model.