The online labour market's expansion presents unique opportunities to analyse job trends through Machine Learning (ML). However, the effectiveness of ML depends on access to well-labelled job advertisement datasets, which are often limited and require labour-intensive manual annotation. Our proposed solution, JobGen, leverages Large Language Models (LLMs) to generate synthetic Online Job Advertisements (OJAs), using real data and the ESCO taxonomy to ensure accurate representation of job market distributions. JobGen enhances data diversity and semantic alignment, addressing common issues in synthetic data generation. The resulting dataset, JobSet, provides a valuable resource for tasks like skill extraction and job matching and is openly available to the community.