Chanwoo Lee, Youngeun Seo, Hyojin Lim, Yum Jung, Haeun Kim, Cheoneum Park, Jisook Park, Journal of Internet Computing and Services, Vol. 27, No. 1, pp. 221-230, Feb. 2026
10.7472/jksii.2026.27.1.221, Full Text:  HTML
Keywords: College Admission Interview, AI-based Mock Interview System, layered prompting, LLM(Large Language Model)

Abstract

This paper proposes Fiterview, an AI-based mock interview system designed to alleviate regional and economic disparities in the college admission interview preparation process. Based on school records and department information, Fiterview automatically generates interview questions using layered and few-shot prompting techniques, and minimizes latency in the follow-up question generation and answer analysis processes through a multi-agent architecture combined with a FastAPI–Celery–Redis-based asynchronous processing structure. For user responses, it provides quantitative evaluations of non-verbal and verbal habits, along with qualitative feedback through LLM-based analysis of the answer content. Experimental results showed that Fiterview demonstrated practical learning effectiveness and feedback utility for interview preparation. Ultimately, this research contributes to fostering educational equity and efficiency by offering an accessible, personalized training platform that reduces the need for private education in interview preparation.

Statistics
Cite this article