CONFERENCE / ICCAIS-2026

Technical interviews constitute a critical component of software engineering recruitment, yet candidates frequently lack access to realistic practice opportunities that simulate authentic interview dynamics. This paper presents a novel web-based platform that integrates OpenAI’s GPT-4o large language model with real-time speech recognition and synthesis capa-bilities, coupled with a professional code editor environment. The system enables seamless voice-to-voice interaction between candidates and an AI interviewer while maintaining real-time code context awareness. Our implementation achieves sub-3-second response latency with 90-95% speech recognition accuracy and provides immediate, context-aware feedback during multi-round interview simulations. Performance evaluation demon-strates successful integration of natural language processing, speech technologies, and code execution in a unified platform, offering an accessible 24/7 interview practice solution.