QuietPrep — helping my friend Nilesh approach practice with more confidence

This submission for Hacktoberfest Weekend Challenge focuses on a personal project to help a friend, Nilesh, prepare for the GATE exam with increased confidence. Nilesh, like many others, grappled with the fear of rejection while studying. QuietPrep, a private practice partner designed for laptop use, offered Nilesh interview-style questions and guidance on explaining technical thinking.

It encourages answering, reflecting, and retrying questions to foster learning. The platform does not measure GATE syllabus coverage or mock-exam scoring, as those aspects are yet to be incorporated. The user experience is simple: select a role and practice focus, answer one question, read one coaching suggestion, retry the same question, and compare attempts.

A story helper organizes a real-life experience around three questions: what needed fixing, personal actions taken, and outcomes learned. After each retry, the system highlights added and removed words, allowing the learner to review changes. An optional AI reflection interprets the revision using passages from both attempts. QuietPrep runs locally on a laptop, with Python's standard library and browser JavaScript powering its operations.

An MIT license governs the source code, while model and runtime licenses are separately documented. Development began on October 3, 2026, using Qwen3-4B-Instruct-2507, with a smaller Qwen2.5-1.5B option for laptops with limited memory. The model generates questions, reviews answers, and provides reflections. The system extracts numbered passages from answers, allowing the model to select a passage ID and display the original text.

Comparisons between attempts help identify where quotations originated. Despite the lack of a countdown or employability score, the learner benefits from ample time to attempt answers and decide on improvements. The evaluation includes failures and snapshots of earlier responses, along with twenty Python tests and four JavaScript tests.

QuietPrep works offline and without an internet connection or cloud API account, making it a locally-controlled learning tool. The compact option, with less memory usage and faster runtime, is available for those with memory constraints. Nilesh has experienced a positive change, feeling less afraid of the exam after using the platform.

Moving forward, the goal is to understand which interactions have helped Nilesh and refine the tool accordingly.