A Semantic Approach for Automated Hiring using Artificial Intelligence & Computer Vision

This research paper presents an automated interview system that utilizes natural language processing and deep learning techniques to conduct job interviews. The system comprises 4 modules, the first one being resume classification system, which is built on state-of-the-art NLP architecture and transformer-based BERT model. The model uses Named Entity Recognition and Text Classification to extract relevant information from the resume. The resume classification algorithm is based on the principle of context relevance, and skill mentions in the resumes are validated through an analysis of the candidate's experience and completed projects before being classified. The subsequent module utilizes the output from the resume model to generate a set of relevant interview questions for the candidate through utilizing natural language processing techniques and semantic analysis. The difficulty of the following questions is adjusted owing to the interviewee’s answer, very much like what human HRs do. Such questions are extracted with the help of web scraping by selenium driver. One of the most innovative parts of this system is its ability to reduce bias in the hiring process by using machine learning models to evaluate candidate responses. The following module assesses the candidate's personality during the interview using the CNN and LSTM models. In addition to generating relevant interview questions, the system also detects potential instances of misconduct or malpractice during the interview process using Open CV’S DNN module and a quantized model. Overall, this system has the potential to revolutionize the way companies conduct job interviews and improve the hiring process.