Fair and Scalable Resume Screening Using Generative AI

Abstract

The increasing volume of applications per job posting has made manual resume screening inefficient, error-prone, and susceptible to unconscious bias. This paper presents a scalable, bias-aware resume screening framework powered by the ChatGPT-4o Large Language Model. The proposed system automatically extracts and weights key skills and qualifications from job descriptions, analyzes candidate resumes in conjunction with publicly available online profiles (e.g., GitHub, LinkedIn), and ranks applicants based on semantic alignment with job requirements. To ensure fairness, the system anonymizes bias-sensitive attributes such as gender, age, race, and disability prior to analysis. A prompt-engineered pipeline guides the model through structured evaluation steps, providing interpretable ranking rationales. The approach improves the consistency, transparency, and equity of candidate selection, while reducing the overall time to screen resumes and offering a practical solution for modern recruitment challenges.

Ähnliche Arbeiten