AI Learning Coach

1) System Architecture

AI Learning Coach is a full-stack learning platform for a 90-day programming bootcamp.

• Frontend (React + Vite)
  • SPA with React Router pages: Dashboard, Lesson, Assignment, Progress, Weekly Challenge.
  • Axios service layer calls backend APIs.
  • Monaco Editor provides Python coding experience.
• Backend (Node.js + Express)
  • REST API for lessons, assignments, progress, and reflections.
  • Mongoose models persist users, progress, assignments, and reflections in MongoDB.
  • Gemini service generates lessons and evaluates Python submissions.
• Database (MongoDB)
  • Stores persistent progress and learning artifacts.
• AI Layer (Google Gemini API)
  • Dynamic lesson generation and code evaluation.
  • Includes fallback behavior if API key is unavailable.

2) Folder Structure

root/
├── client/
│   ├── src/
│   │   ├── components/
│   │   │   ├── Navbar.jsx
│   │   │   ├── ProgressBar.jsx
│   │   │   └── CodeEditor.jsx
│   │   ├── pages/
│   │   │   ├── Dashboard.jsx
│   │   │   ├── Lesson.jsx
│   │   │   ├── Assignment.jsx
│   │   │   ├── Progress.jsx
│   │   │   └── WeeklyChallenge.jsx
│   │   ├── services/
│   │   │   └── api.js
│   │   ├── styles/
│   │   │   └── main.css
│   │   ├── App.jsx
│   │   └── main.jsx
│   ├── index.html
│   ├── package.json
│   └── vite.config.js
├── server/
│   ├── controllers/
│   │   ├── lessonController.js
│   │   └── assignmentController.js
│   ├── routes/
│   │   ├── lessonRoutes.js
│   │   └── assignmentRoutes.js
│   ├── models/
│   │   ├── User.js
│   │   ├── Progress.js
│   │   ├── Assignment.js
│   │   └── Reflection.js
│   ├── services/
│   │   └── geminiService.js
│   ├── .env.example
│   ├── package.json
│   └── server.js
└── README.md

3) Installation Steps

Prerequisites

• Node.js 18+
• MongoDB running locally (or cloud URI)
• Gemini API key

Backend

cd server
npm install
cp .env.example .env
# edit .env with your MONGO_URI and GEMINI_API_KEY
npm run start

Frontend

cd client
npm install
npm run dev

Open http://localhost:5173.

API Endpoints

• GET /lesson/:day
• POST /assignment/submit
• GET /progress
• POST /reflection
• GET /dashboard

Troubleshooting

If backend start fails with EADDRINUSE: address already in use :::5000:

• Either stop the process using that port:
  • macOS/Linux: lsof -i :5000 then kill -9 <PID>
• Or start on a different preferred port:
  • PORT=5001 npm run start

The backend automatically retries with the next free port when the preferred port is occupied. By default it retries up to 15 additional ports (PORT_RETRY_LIMIT=15). You can increase this range if needed:

• PORT=5000 PORT_RETRY_LIMIT=30 npm run start