FairHire: A Fair and Automated Candidate Screening System

The evolution of online job postings has drastically changed the recruitment landscape, inundating Human Resource (HR) teams with an overwhelming influx of applicants. Addressing this challenge, this paper presents a web-based application automating candidate screening and streamlining the talent acquisition process by leveraging Machine Learning (ML) and Large Language Models (LLM). The system employs a duallayered screening process: a Naive Bayes Classifier filters resumes by job categories, followed by Natural Language Processing techniques and cosine similarity to rank resumes against job descriptions. Following the generation of ranked list, HR can engage with shortlisted individuals via integrated email service facilitated by RabbitMQ microservice. Additionally, the system integrates an LLM based on GPT-2, to generate interview questions aligned with candidates’ extracted skills from their resumes. This paper delineates a structured approach to automate candidate screening processes highlighting the potential impact of this innovative approach on HR practices, fostering efficiency in candidate evaluation.