Artificial Intelligence (AI)-Based Resume Analysis and Job Matching

The transition from academic education to professional employment requires students to communicate technical skills, projects, certifications, education, and experience in a form that is relevant to a specific job. This paper presents an AI-based resume analysis and job-matching approach designed to provide understandable, job-specific feedback to engineering students. The system accepts a PDF resume and a target job description, extracts resume text using pdfplumber with pypdf as a fallback, and submits the normalized information to an AI analysis component. The revised methodology makes the compatibility score more transparent by interpreting it through three dimensions: required-skill coverage, semantic relevance between resume evidence and job requirements, and evidence strength from projects, education, and experience. The system returns four structured outputs: an ATS-style compatibility score, matched skills, missing or insufficiently represented skills, and actionable recommendations. Unlike a commercial Applicant Tracking System, the score is explicitly positioned as a decision-support estimate rather than an employer-generated recruitment score. Implementation-level testing on three sample technical cases produced compatibility scores of 78%, 64%, and 71%, with a mean of 71.0%. These values demonstrate functional variation across job-specific inputs but are not treated as predictive accuracy because the current study does not use a sufficiently large labeled benchmark. A literature-grounded comparison further shows that existing approaches emphasize resume information extraction, skill taxonomy alignment, deep-learning recommendation, or job recommendation, whereas the proposed system integrates these ideas into a student-facing feedback workflow. The paper also identifies limitations related to dataset size, PDF extraction, AI variability, fairness, and external validation, and defines a future research plan involving a larger annotated benchmark, human agreement analysis, taxonomy-based skill normalization, and systematic evaluation against baseline matching methods.