Job seekers primarily aim to find opportunities that align with their skills and qualifications, while human resource (HR) departments strive to identify candidates whose resumes meet employers’ criteria. Given the vast number of job seekers and openings, manually reviewing resumes and listings has become impractical. To address this, recommender systems have been developed to assist both job seekers and HR departments by suggesting suitable matches. However, many existing systems focus on broad classification techniques that lack the specificity required for accurate job matching. In this work, we propose LLM4RecSys, a novel recruitment recommender system that leverages large language models (LLMs) to understand and match job offers with resumes in a context-aware manner. Unlike traditional classification approaches, LLM4RecSys is tuned to capture the intricate relationships between candidate qualifications and job requirements. By analyzing both resumes and job descriptions in detail, the system provides tailored, relevant recommendations. Experiments on an open dataset of resumes and job offers demonstrate that LLM4RecSys significantly improves matching accuracy over traditional deep learning models such as BERT. Additionally, the model enhances recruiters’ decision-making by offering clearer explanations for its recommendations, enabling better-informed hiring choices.