JOBSKAPE: A Framework for Generating Synthetic Job Postings to Enhance Skill Matching

Recent approaches in skill matching, employing synthetic training data forclassification or similarity model training, have shown promising results,reducing the need for time-consuming and expensive annotations. However,previous synthetic datasets have limitations, such as featuring only one skillper sentence and generally comprising short sentences. In this paper, weintroduce JobSkape, a framework to generate synthetic data that tackles theselimitations, specifically designed to enhance skill-to-taxonomy matching.Within this framework, we create SkillSkape, a comprehensive open-sourcesynthetic dataset of job postings tailored for skill-matching tasks. Weintroduce several offline metrics that show that our dataset resemblesreal-world data. Additionally, we present a multi-step pipeline for skillextraction and matching tasks using large language models (LLMs), benchmarkingagainst known supervised methodologies. We outline that the downstreamevaluation results on real-world data can beat baselines, underscoring itsefficacy and adaptability.