VOICE-ENABLED SELF INTERVIEW EVALUATION SYSTEM USING AI AND NLP

Abstract

With the increasing demand for intelligent and scalable interview preparation systems, traditional methods fail to provide personalized, unbiased, and real-time evaluation. This paper presents a Voice-Enabled Self Interview Evaluation System that integrates Artificial Intelligence (AI), Natural Language Processing (NLP), and speech technologies to simulate realworld interview environments. The proposed system allows users to answer interview questions through both text and voice inputs, enabling a more realistic experience. The system leverages Sentence-BERT (SBERT) for semantic similarity evaluation, combined with keyword matching and answer length analysis to generate accurate scores and feedback. Additionally, resume-based personalization extracts candidate skills to tailor interview questions dynamically. A MongoDBbased backend supports dynamic question retrieval, including web-scraped data, while a Flask-based architecture ensures scalability and modularity. Experimental analysis shows that the system provides reliable evaluation, reduces human bias, and improves user engagement through real-time feedback and performance analytics. The proposed solution offers a comprehensive, intelligent, and user-friendly platform for modern interview preparation