A Review on AI-Powered SmartInterview System Using BERT-CNN for Semantic Answer

The proposed AI Interview Assistant with Voice, Emotion and Skill Analysis is an intelligent Android application developed to help candidates prepare for job interviews through automated assessment and personalized feedback. Traditional interview preparation methods generally depend on self-practice, coaching, or guidance from experienced individuals, which may not provide continuous performance evaluation. To overcome these limitations, the proposed system combines Artificial Intelligence, Machine Learning, Computer Vision, Speech Processing and Natural Language Processing to analyse different aspects of interview performance. A CNN-based model is used to identify facial expressions, while MFCC-based feature extraction and machine-learning techniques are applied to analyse voice characteristics and speech emotions. Speech-to-text technology converts spoken responses into text, which is further processed using NLP techniques to examine answer relevance, important keywords, communication quality and technical skill alignment. The results obtained from facial, voice and textual analysis are combined to generate an overall performance report. The system also provides personalized recommendations, interview improvement tips, learning resources and skill development suggestions. Firebase is used to maintain interview history and monitor performance over multiple practice sessions. The proposed application aims to provide an accessible and interactive interview preparation platform that supports candidates in improving their communication, technical knowledge and behavioural presentation skills.