Nemo

AI-powered software interview preparation. Students get a personalised day-by-day prep plan, complete tasks verified by AI-generated MCQs, earn diamonds, unlock companions, and track progress through a pixel-art character that evolves with their mood.

Features

• Deterministic stage-based prep plan.
• AI-generated MCQ validation via Amazon Nova (Amazon Bedrock).
• Pixel art character with gender selection and mood states.
• Pokemon companions from shop.
• Streak and diamond reward system.
• Badge achievements.
• Calendar view of full prep timeline.
• Adaptive replan placeholder.

Tech Stack

• React + Vite.
• TypeScript.
• TailwindCSS.
• Supabase (PostgreSQL).
• Amazon Nova Lite (Amazon Bedrock).
• FastAPI Python backend.

Architecture

Frontend (React/Vite) -> Supabase (auth + database) -> FastAPI backend -> Amazon Nova Lite via Amazon Bedrock for MCQ generation.

AI Integration

When a user completes a task, a live Amazon Nova Lite call (via Amazon Bedrock) generates a unique MCQ specific to the topic and difficulty. No two attempts see the same question.

How to Run Locally

• Clone the repo.
• Run npm install.
• Create .env in the root folder using .env.example as reference and fill in your keys.
• Run npm run dev for frontend.
• Run uvicorn main:app --reload inside nemo-backend for backend.