Automated Resume Screening And Segmentation Using Natural Language Processing

Abstract

An AI-based system designed to simplify and improve the recruitment process through Artificial Intelligence and Natural Language Processing (NLP) techniques. The system automatically extracts information from PDF resumes and compares it with a job description to identify the most suitable candidates. It uses Term Frequency-Inverse Document Frequency (TF-IDF) for keyword extraction, Sentence-BERT (SBERT) for semantic similarity analysis, and Applicant Tracking System (ATS) scoring for keyword-based evaluation. A weighted scoring mechanism combines these measures to generate a final candidate score and rank applicants accordingly. The system further segments resume content into sections such as skills, education, experience, projects, and certifications, and includes an Explainable AI module that highlights missing skills and gives transparent reasons for each candidate score. An AI Resume Suggestion module recommends improvements to resume quality and ATS compatibility, while a Resume Quality Checker verifies the completeness of essential details such as contact information, skills, projects, LinkedIn, GitHub, and certifications. A classification model further predicts the most suitable job role from resume content. All outputs are presented through an interactive Streamlit dashboard showing candidate rankings, performance metrics, and graphical visualizations. The prototype was implemented using Python, Streamlit, spaCy, Sentence-Transformers, Scikit-learn, Pandas, NumPy, Matplotlib, and pdf plumber, and testing confirms that it reduces manual screening effort, improves ranking accuracy over keyword-only ATS filtering, and supports fair, transparent, and data-driven hiring decisions.