Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{202849,
        author = {Miss Priyanka Mohan Doijode and Dr.Sushilkumar N Holambe and Kishorekumar Vishwanathan},
        title = {InterviewAI: A Voice-Driven Multi-Agent AI Framework for Adaptive Mock Interviews with Real-Time Feedback},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {9283-9290},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=202849},
        abstract = {Interview preparation remains a critical challenge for job seekers due to the absence of scalable, realistic, and feedback-driven practice environments. Conventional mock interview platforms rely heavily on scripted text interfaces or static question banks, limiting their ability to evaluate real conversational competence, vocal confidence, and contextual reasoning. This paper presents Prepwise, a voice-driven artificial intelligence interview preparation platform that integrates real-time conversational agents, adaptive question generation, and automated performance analytics within a cloud-native architecture. The proposed system leverages Vapi AI voice agents for natural spoken interaction, Google Gemini for dynamic interview question synthesis, and Firebase for secure authentication, session persistence, and interview lifecycle management. A modular multi-agent workflow is designed to coordinate interviewer behavior, response evaluation, transcript generation, and structured feedback scoring. Unlike traditional systems, InterviewAI continuously adapts interview difficulty and topic progression based on candidate responses, semantic coherence, and detected hesitation patterns. Post-interview, the platform generates granular feedback across technical accuracy, communication clarity, confidence indicators, and response relevance using structured prompt-engineering and rubric-based scoring models. Experimental deployment demonstrates that the system achieves low-latency voice interaction, consistent interview reproducibility, and reliable feedback alignment across technical and behavioral domains. The architecture emphasizes scalability, device independence, and rapid extensibility to multiple job roles without retraining core models. The proposed framework establishes a novel direction for AI-mediated skill assessment by unifying voice-centric interaction, adaptive reasoning, and automated qualitative evaluation into a single interview intelligence pipeline.},
        keywords = {Artificial intelligence interview systems, conversational agents, voice-based human–AI interaction, automated feedback generation, adaptive assessment platforms.},
        month = {May},
        }