AI Interview Preparation Environment

Production-ready RL system for technical interview practice with intelligent feedback and OpenEnv compliance.

🚀 Quick Start

Option 1: Docker (Recommended)

docker build -t ai-interview-env .
docker run -p 7860:7860 -e HF_TOKEN=your_token ai-interview-env

Option 2: Local Development

Backend:

cd backend
source venv/bin/activate  # or: venv\Scripts\activate on Windows
pip install -r requirements.txt
python main.py

Frontend:

cd frontend
npm install
npm run dev

Access at http://localhost:3000

Option 3: Run Inference (OpenEnv)

python3 inference.py

⚙️ Configuration

Environment Variables

export API_BASE_URL=https://router.huggingface.co/v1
export MODEL_NAME=Qwen/Qwen3-Coder-Next:novita
export HF_TOKEN=your_huggingface_token  # Optional, uses mock mode if not set

UI Configuration

• Click ⚙️ button in top-right
• Enter your HuggingFace API token
• Select or enter model name
• Save configuration

Supported Models

• Qwen/Qwen3-Coder-Next:novita (default)
• meta-llama/Llama-3.3-70B-Instruct
• Qwen/Qwen2.5-72B-Instruct
• mistralai/Mixtral-8x7B-Instruct-v0.1
• Any HuggingFace model with inference API

📚 Features

• 12 Technical Questions: Algorithms, systems, databases, OOP, design patterns
• Deterministic Grading: Keyword matching (80%) + length scoring (20%)
• Structured Feedback: Missing & matched keywords identified
• Retry Mechanism: Agent improves answers using feedback
• Memory System: Tracks attempts and improvement
• OpenEnv Compliant: Full compliance with OpenEnv hackathon requirements
• Dynamic Configuration: Change models without restart
• Beautiful UI: React interface with real-time updates

🏗️ Project Structure

ai-interview-env/
├── openenv.yaml        # OpenEnv specification
├── inference.py        # Inference script
├── Dockerfile          # Container configuration
├── backend/
│   ├── app/
│   │   ├── agent.py      # LLM agent with retry
│   │   ├── environment.py # RL environment
│   │   ├── evaluator.py  # Deterministic scoring
│   │   └── dataset.json  # 12 questions
│   ├── main.py           # FastAPI server
│   └── requirements.txt
├── frontend/             # React application
└── docs/                 # Documentation
    ├── OPENENV.md        # OpenEnv compliance details
    └── PROJECT.md        # Full architecture guide

🔌 API Endpoints

OpenEnv Endpoints

• POST /reset - Reset environment, get new question
• POST /step - Submit answer, get reward
• GET /state - Get current state

Legacy Endpoints (for frontend)

• GET /question - Get new question
• POST /answer - Submit answer
• POST /auto-run - Auto-generate & improve
• GET /config - Get current config
• POST /config - Update API key/model
• GET /stats - Episode statistics

✅ OpenEnv Compliance

This project is fully compliant with OpenEnv hackathon requirements:

• ✅ openenv.yaml specification
• ✅ Deterministic grading (0.0-1.0 rewards)
• ✅ 12 tasks (≥3 required)
• ✅ Docker containerization
• ✅ /reset, /step, /state endpoints
• ✅ [START]/[STEP]/[END] logging in inference.py

See docs/OPENENV.md for detailed compliance information.

🛠️ Development

Add Questions

Edit backend/app/dataset.json:

{
  "question": "Your question?",
  "keywords": ["key1", "key2"],
  "difficulty": "medium"
}

Test API Locally

# Start server
cd backend && python3 main.py

# Test in another terminal
curl -X POST http://localhost:8000/reset
curl -X POST http://localhost:8000/step \
  -H "Content-Type: application/json" \
  -d '{"action":"Binary search divides a sorted array"}'
curl http://localhost:8000/state

Run Tests

cd backend
pytest

📖 Documentation

• docs/OPENENV.md - OpenEnv compliance and validation
• docs/PROJECT.md - Complete architecture details and development guide

📄 License

MIT License - Free for educational and commercial use

🙏 Credits

Built for OpenEnv Hackathon 2026 with FastAPI, React, and HuggingFace models.