LLM-Based Interview Assistance and Q&A Generator

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Preparing for technical interviews requires repeated practice, relevant questions, and constructive feedback.Learners often rely on static question banks or informal practice, which may not adapt to a selected role, topic, or the learner's response.This paper presents the design of an LLM-based interview assistance and question-and-answer (Q&A) generation system intended to support structured interview practice and, in a planned extension, quiz generation.The current prototype provides a web-based interaction flow using Flask, Google Gemini through LangChain, AssemblyAI speech-to-text, and Murf text-to-speech.A response can be captured, processed, and passed to a language model for contextual feedback.The proposed next phase extends the same foundation with Retrieval-Augmented Generation (RAG), which will retrieve relevant material from a curated knowledge base and use it to ground interview questions, model answers, explanations, and quiz items.The quiz-generation module and RAG integration are planned extensions and are not reported as completed implementation.Since formal user studies and quantitative benchmarking have not yet been conducted, this paper describes the system design, expected workflow, research motivation, and evaluation plan rather than claiming measured performance.The system is intended as a learning aid, not as a replacement for human interviewers or expert review.