Generative AI in HRM : exploring fine-tuning large language models for candidate sourcing and screening

Abstract

This Thesis explores the implementation of Gen AI, particularly Large Language Models in the process of transforming traditional Human Resource Management functions like candidate sourcing and candidate screening respectively. As the organizations are currently implementing AI technologies in different operations, HRM is evolving from traditional admin and human oriented functions to more strategic based AI driven solutions significantly improving recruitment efficiency and improving quality of hiring candidates. This study adopts a SLR approach examining the research publications from the year 2018 to 2025, covering both industry and academic perspectives. It evaluates how general-purpose and fine-tuned LLMs are utilized across functions like resume screening, job matching, and evaluation of a job applicant.

The analysis highlights differences in the performance, reliability and accuracy between baseline models and domain-specific tuned systems in different factors like bias, fairness and interpretability. Overall, results suggest that generative AI can significantly improve recruitment process when applied with appropriate fine tuning and human interference. The study emphasizes that LLMs should function as assistive tools rather than autonomous decision makers.