---

# Offline-First AI Tutor for GCE Students  
**Group 11 – SEED Inc. Innovation Program**  
**Team Members:** Tayuh Favour (PM) • Nditafon Grace • Njomen Jeanson • Sale Aieshatou • Neba Stanislas 

---

## 🧠 Project Overview
The **Offline-First AI Tutor** is a web-based learning assistant built to support **Cameroon GCE students** who face poor internet connectivity and limited access to quality tutoring.  
The system uses **Lightweight LLMs** (via Ollama) and **Retrieval-Augmented Generation (RAG)** to provide curriculum-aligned, interactive explanations — even offline.

---

## 🎯 Problem Statement
Many students preparing for the **Cameroon GCE examinations** struggle due to:
- Unreliable or costly internet access.
- Limited access to qualified tutors.
- Lack of personalized study guidance and timely feedback.

This project bridges the **urban-rural educational gap** by delivering AI-driven learning support that runs on low-end devices and functions offline.

---

## 💡 Proposed Solution
A web-based **Offline-First AI Tutor** that:
- Uses an **Ollama-hosted local model** for subject-specific Q&A (Mathematics, Physics, Biology, etc.)
- Implements **RAG** using a locally indexed database of GCE past papers, marking guides, and textbooks.
- Works **offline-first**, syncing progress and updates whenever an internet connection is detected.
- Provides **personalized learning paths** with analytics, revision plans, and topic recommendations.

---

## 🧰 Tech Stack

| Layer | Technology | Purpose |
|-------|-------------|----------|
| **Frontend** | Next.js | Web interface for students and teachers |
| **Backend** | FastAPI | Handles AI requests, RAG pipeline, and database communication |
| **AI Model** | Ollama (Phi / Mistral) + LangChain | Local lightweight LLM for explanations and RAG |
| **Database** | SQLite (offline), Supabase (cloud sync) | Store user data, progress, and GCE content |
| **Hosting / Deployment** | Render / Supabase Edge | For online sync and updates |

---

## ⚙️ Folder Structure

ai-tutor-gce/ ├── frontend/ # Next.js web app (student interface) │ ├── components/ │ ├── pages/ │ ├── public/ │ └── utils/ ├── backend/ # FastAPI server and API endpoints │ ├── app/ │ │ ├── main.py │ │ ├── routes/ │ │ ├── services/ │ │ └── models/ │ └── requirements.txt ├── model/ # Ollama integration, fine-tuning scripts │ ├── load_model.py │ └── rag_pipeline.py ├── data/ # GCE dataset and past papers │ ├── raw/ │ └── processed/ ├── docs/ # Documentation (SRS, diagrams, reports) ├── .gitignore ├── README.md └── LICENSE

---

## 🔄 Collaboration Workflow

1. **Main branch:** Stable, production-ready code only.
2. **Feature branches:** Each team member works on their assigned branch.
   - Example:  
     - `frontend-grace` → UI/UX  
     - `backend-jeanson` → API & RAG integration  
     - `model-favour` → Ollama + LangChain setup  
     - `docs-group11` → Reports, diagrams, and SRS
3. Pull Requests → Code Reviews → Merge to `main`

### 🧩 Branch Naming Convention
| Branch Type | Format | Example |
|--------------|--------|----------|
| Frontend | `frontend-<name>` | `frontend-grace` |
| Backend | `backend-<name>` | `backend-jeanson` |
| AI Model | `model-<name>` | `model-favour` |
| Documentation | `docs-<name>` | `docs-group11` |

---

## 🧪 Setup Instructions

### 1️⃣ Clone Repository
```bash
git clone https://github.com/<your-username>/ai-tutor-gce.git
cd ai-tutor-gce

2️⃣ Frontend Setup

cd frontend
npm install
npm run dev

3️⃣ Backend Setup

cd backend
pip install -r requirements.txt
uvicorn app.main:app --reload

4️⃣ Model Setup (Ollama)

Install Ollama locally: https://ollama.ai/download

Then pull a lightweight model:

ollama pull phi

Run it:

ollama run phi

---

📅 Project Timeline (3 Weeks)

Week	Milestone	Deliverable
Week 1	Requirement analysis & setup	SRS, folder structure, diagrams
Week 2	Core development	Frontend + Backend + AI integration
Week 3	Testing & refinement	Final demo, documentation, submission

---

👥 Roles and Responsibilities

Member	Role	Responsibility
Tayuh Favour	Project Manager / AI Developer	Lead dev workflow, model integration, review PRs
Nditafon Grace	Frontend Developer	Build student dashboard, UI design
Njomen Jeanson	Backend Engineer	Build APIs, manage RAG, connect DBs

---

📦 Collaboration Tools

Tool	Purpose
GitHub	Version control & code collaboration
Trello	Project management (3-week roadmap)
Google Docs	Shared notes & reports
Draw.io	Diagrams and architecture design
Supabase Dashboard	Cloud database management

---

🧩 Contribution Guide

1. Create a branch for your task.
2. Make your changes, commit often with clear messages.
3. Push your branch and open a Pull Request.
4. Tag the PM for review.
5. Wait for approval and merge.

---

📜 License

This project is licensed under the MIT License — free to use, modify, and distribute with attribution.

---

📧 Contact

SEED Inc. – Nurturing the Next Generation of Tech Talent Center Bolt, Mile 4, Bamenda, Cameroon 📩 Email: [email protected] | ☎️ +237 680 468 606 🌐 Facebook | Instagram | LinkedIn

---