AHGV Local AI Tutor：注重私隱的 AI 學習助理

AHGV Local AI Tutor: A Privacy-Focused AI Study Assistant 🎓

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle.

What I Built

AHGV Local AI Tutor is an AI-powered study assistant designed to help students learn from their own study materials instead of repeatedly searching through lengthy documents.

Students can provide PDF, Markdown, and text files, or paste their notes directly. The application helps turn those materials into useful learning resources.

Key features

Smart summaries: Convert lengthy study materials into concise summaries.
Revision notes: Generate structured notes for exam preparation.
MCQ generation: Create multiple-choice questions for self-assessment.
Context-aware Q&A: Ask questions about uploaded study materials and receive answers grounded in relevant content.
Lightweight retrieval: Find relevant passages using lexical retrieval before generating answers.
Student-focused workflow: Bring notes, explanations, and revision tools into one place.

The intended users are college students, independent learners, and tutors who need a practical way to organize study material and prepare for examinations.

The long-term goal is to make useful AI-assisted learning accessible without requiring students to maintain expensive hardware or purchase multiple study tools.

Demo

🌐 Project: https://ahgv-learning.vercel.app

The companion project is available online. The AHGV Local AI Tutor repository contains the code for the study-assistant implementation.

Code

💻 GitHub Repository: https://github.com/Shivam-Sharma2009/AHGV_PROJ

The project uses a React and Vite frontend with a Python FastAPI backend. Supporting technologies include Pydantic, HTTPX, PyPDF, and pytest for validation and testing.

I am building this as a practical learning project, with an emphasis on keeping the architecture lightweight and making the code understandable and extensible.

How I Built It

I built AHGV Local AI Tutor around a simple workflow:

Input: Accept study material from PDF, Markdown, text files, or pasted notes.
Processing: Extract and prepare the text for retrieval.
Retrieval: Identify relevant passages using lexical matching.
AI assistance: Use the retrieved context to support summaries, revision material, quizzes, and question answering.
Output: Present the results in a student-friendly interface.

The frontend uses React 19 and Vite. The backend uses FastAPI and Uvicorn, with Pydantic for data validation, HTTPX for HTTP communication, and PyPDF for PDF processing.

My goal is to support open-weight AI models while keeping the application practical on limited hardware. Because my laptop has limited storage, running a local model through Ollama has been a constraint. I am exploring lightweight deployment options rather than requiring every student to download a large model.

The next step is to make the open-weight inference path fully reproducible and document the model and deployment requirements clearly.

Why Does Open Innovation Matter?

Open innovation makes AI development more accessible to students and independent developers who do not have access to expensive GPUs or large cloud budgets.

For a project like AHGV Local AI Tutor, open-weight models can provide greater control over model selection, deployment, and customization. They