AI Resume Screening and Job Description Matching System

Abstract

The modern recruitment landscape is overwhelmed by the sheer volume of job applications received for every available position, creating significant bottlenecks in the hiring process. Human recruiters spend a disproportionate amount of time manually reviewing resumes, leading to delays, inconsistencies, and potential bias in candidate evaluation. This paper presents the design, development, and evaluation of an AIpowered Resume Screening and Job Description Matching System that leverages Natural Language Processing (NLP) and machine learning-based text similarity techniques to automate candidate evaluation. The proposed system accepts candidate resumes in PDF format, extracts technical skills, qualifications, certifications, and experience using an NLP pipeline built on NLTK and spaCy, and compares the extracted information against job description requirements through keyword matching and TF-IDF cosine similarity scoring via Scikit-learn. The backend is developed using Python and Flask, while the frontend is constructed with HTML5, CSS3, and JavaScript, delivering a responsive and recruiter-friendly web interface. The system generates a quantitative skill match percentage alongside detailed insights into matched and missing skills, enabling data-driven hiring decisions. Experimental testing across multiple job profiles and resume formats demonstrated reliable skill extraction accuracy and match percentage consistency. The modular system architecture supports future enhancements including transformer-based semantic matching, multi-resume batch processing, AI-based candidate ranking, and cloud deployment. This work contributes a practical, scalable, and deployable solution to the growing field of intelligent recruitment and talent acquisition automation.

Keywords