Saarthi AI (HACKQUINOX WINNER)

🚀 Overview

AI Interview Coach is an advanced AI-powered interview simulator that helps users prepare for job interviews by mimicking real-life interview scenarios. It uses speech-to-text (STT), text-to-speech (TTS), NLP-based question generation, and facial emotion recognition (FER) to evaluate responses and provide insightful feedback.

🔥 Features

• Real-time Speech-to-Text (STT): Uses Web Speech API to transcribe user responses.
• Text-to-Speech (TTS): Converts AI-generated interviewer questions into natural speech using ElevenLabs AI voices.
• Dynamic Question Generation: Generates context-aware interview questions using Gemini AI.
• Facial Emotion Recognition (FER): Analyzes user expressions using DeepFace / FER+ models.
• Real-time Feedback: Provides spoken and visual feedback based on voice tone and facial expressions.
• Interactive Visualization: Uses a custom audio player with waveform visualization instead of a basic HTML <audio> component.
• Multi-role Support: Users can select different interview types (e.g., Data Analyst, Software Engineer, Product Manager).

🛠️ Tech Stack

• Frontend: React.js, Tailwind CSS, Web Speech API
• Backend: Node.js, Express.js
• AI Models:
  • Speech Recognition: Web Speech API
  • Text-to-Speech: ElevenLabs AI
  • Question Generation: Google Gemini AI
  • Emotion Detection: DeepFace / FER+ (Microsoft)
  • Face Tracking: OpenCV
• Database: MongoDB (Optional for storing history and feedback)

⚙️ Installation & Setup

1️⃣ Clone the Repository

git clone https://github.com/your-username/ai-interview-coach.git

2️⃣ Install Dependencies

# Frontend
cd client
npm run dev

3️⃣ Set Up API Keys

Create a .env file in the server directory and add:

ELEVEN_LABS_API_KEY=your-elevenlabs-api-key
GEMINI_API_KEY=your-gemini-api-key

4️⃣ Start the Application

# Start backend
cd server
node index.js

📌 Usage

1. Click Start Talking to begin the interview.
2. The AI will ask a question using TTS.
3. User responds, and speech is transcribed to text.
4. AI analyzes the response and provides feedback on:
   • Clarity
   • Emotion
   • Relevance
5. Repeat for multiple rounds.

🎯 Future Improvements

• ✅ Multi-language support for global users.
• ✅ More advanced interviewer personalities (strict, friendly, technical).
• ✅ Integration with job portals to analyze real interview trends.

🤝 Contributing

We welcome contributions! Feel free to submit issues or pull requests.

📜 License

This project is licensed under the MIT License.

---

✨ Created by [SUJIT_MISHRA / DEV_PIONEERS] 🚀