AI-Powered Interview Simulator: A Web-Based Platform for Personalised Interview Practice with ATS Resume Analysis, Voice and Visual Analysis

Interviews are the decisive stage of recruitment, yet many students and job seekers lack realistic, personalised practice and rarely know how an Applicant Tracking System (ATS) will judge their resume. This paper presents the design of an AI-Powered Interview Simulator, a web-based platform that brings ATS resume analysis, personalised question generation, Technical, Aptitude and HR rounds, voice-based answering and visual behaviour analysis into one application. The uploaded resume is scored using ATS rules, and the Gemini API generates questions from the candidate’s own skills, projects and experience. Spoken answers are converted to text with Whisper and evaluated by the Gemini API for relevance, correctness and clarity, while MediaPipe extracts visual indicators from webcam frames. A scoring module combines these results into an overall score and structured feedback. A survey of twenty-three studies shows that existing tools address only one part of interview preparation. The paper presents the proposed architecture and data flow diagrams that form the basis for implementation.