Published Paper Details:

This paper presents the design, development, and evaluation of an intelligent mock interview platform aimed at improving career readiness through real-time, AI-driven simulations. Leveraging Google's Gemini (PALM) large language model, the system dynamically generates domain-specific interview questions and evaluates user responses based on contextual relevance, fluency, and delivery. Built using JavaScript, Tailwind CSS, Firebase, and WebRTC, the platform delivers an immersive, role-based experience simulating real-world interviews. Users select a job role to receive customized sessions, during which their audio or video responses are analysed in real time. Gemini's semantic understanding enables automated feedback across articulation, content depth, confidence, and role alignment, with session data stored securely in Firebase for review and iterative learning. The architecture comprises a user interface layer, an AI services layer powered by Gemini APIs, and a data services layer. Evaluation with 75 participants from fields such as software engineering, marketing, and data science showed a 38% improvement in performance over three sessions, validated through both automated and expert assessments. User feedback emphasized the platform's realism and adaptability. This study demonstrates the effectiveness of large language models in enhancing communication skills and proposes future enhancements including emotion analysis, multilingual support, and longitudinal impact studies.