International Research Journal of Education and Technology

The growing field of job recommendation systems is increasingly important in today’s dynamic job market. However, existing systems rely heavily on static keyword-based matching, which lacks the ability to accurately capture the intricate relationships between job roles, skills, and qualifications. Therefore, there is a need for a real-time, adaptive system that can address these shortcomings by leveraging advanced techniques. The objective of this research is to develop a real-time job recommendation system that integrates knowledge graphs and user analytics to provide personalized, context-aware job recommendations. The problem lies in current systems’ inability to adapt quickly to changes in user behavior and market trends. To address this, we constructed a dynamic knowledge graph that captures the complex interconnections between job roles, skills, qualifications, and candidate profiles. We employed Graph Neural Networks (GNNs) and graph embedding techniques to process this data. Our methodology involved analyzing real-time user interaction data and job market trends to continuously update the system. The results indicate a significant improvement in job matching accuracy, with a 30% increase in recommendation relevance when compared to static systems. Discussion of the findings shows that the knowledge graph’s continuous updates based on real-time data significantly enhance both the job seeker and employer experience, reducing time-to-hire and improving overall job satisfaction. These results highlight the system’s effectiveness in creating personalized, real-time job recommendations, offering a scalable solution for modern recruitment challenges.

Keywords – job recommendation system, knowledge graph, user analytics, Graph Neural Networks, real-time job matching, graph embedding