Abstract
Technical interview preparation is time-consuming and overwhelming, with candidates navi gating scattered resources and uncertain which questions to practice. This thesis presents JobTutor, a web application that addresses these challenges through two integrated com ponents: a Retrieval-Augmented Generation (RAG) chatbot that provides evidence-based interview guidance from Reddit, Stack Overflow, and GitHub sources, and an optimized question recommendation engine that uses an ILP-inspired greedy algorithm to select ques tions maximizing knowledge component coverage while balancing difficulty for the user’s role and experience level. The RAG chatbot was evaluated using RAGAS, achieving strong performance across key metrics: Answer Relevance (0.8907), Faithfulness (0.8792), Context Recall (0.8273), and Context Precision (0.6564).
Advisor
Visa, Sofia
Department
Computer Science
Recommended Citation
Berhe, Mastewal, "JobTutor: A RAG-Based Chatbot and ILP-Inspired Question Recommender for Technical Interview Preparation" (2026). Senior Independent Study Theses. Paper 13346.
https://openworks.wooster.edu/independentstudy/13346
Disciplines
Computer Engineering
Keywords
RAG, ILP, Knowledge Tracing, Question Recommendation, Technical Interview Preparation, Large Language Models
Publication Date
2026
Degree Granted
Bachelor of Arts
Document Type
Senior Independent Study Thesis
© Copyright 2026 Mastewal Berhe