🤖 AI Interview Bot

An interactive AI Interviewer powered by Google GenAI Live API — conducts real-time, voice-based interviews for any role using your uploaded job description and resume.

---

🧩 Project Overview

This project combines a Python backend (WebSocket + GenAI Live API) with a browser frontend (HTML + JS) to simulate a spoken job interview.

• 🎙️ Live audio input/output
• 📄 Resume parsing from PDF
• 🌍 Multilingual support (EN, TR, RU, AZ)
• 💬 Customizable prompts for different roles
• 🔌 Simple setup — no frameworks required

---

📁 Directory Structure

.
├─ config/                 # Job title & description text files
├─ prompts/                # System prompt template (optional)
├─ ai_interviewer_bot.py   # Python backend server
├─ index.html              # Frontend (runs in browser)
├─ resume.pdf              # Candidate resume
├─ .env                    # API key + project config
├─ .gitignore
└─ README.md

---

⚙️ Prerequisites

• Python 3.10+
• A Google GenAI API key (with Live API access)
• Modern browser with mic + audio support (Chrome, Edge, Firefox)

---

🚀 Setup & Run

1. Clone the repo

git clone <your-repo-url>
cd ai-interview-bot

2. Install dependencies

pip install -U pip
pip install google-genai websockets python-dotenv pypdf

3. Configure your .env

GEMINI_API_KEY=YOUR_API_KEY_HERE
PROJECT_ID=ai_interviewer-bot

4. Add job info & resume

• config/job_title.txt → e.g. AI Engineer
• config/job_description.txt → Paste role details
• resume.pdf → Candidate’s resume

5. Start the backend

python ai_interviewer_bot.py

You’ll see:

Server: ws://localhost:8765
Backend ready! Open index.html in your browser.

6. Open the frontend

Just open index.html in your browser.

---

🌍 Supported Languages

Language	Code
English	en-US
Turkish	tr-TR
Russian	ru-RU
Azerbaijani	az-AZ

---

🎧 How It Works

1. Browser connects via WebSocket → ws://localhost:8765
2. User audio (16 kHz PCM) is streamed to the backend
3. Backend sends it to Google GenAI Live API
4. AI responds with speech + text
5. Browser plays AI’s voice and shows text log

---

🛠️ Customize

• Change number of questions: Edit self.max_questions in InterviewSession (default 6)
• Change voice: Modify voice_name in types.PrebuiltVoiceConfig
• Custom system prompt: Add prompts/system_prompt.txt (uses placeholders {job_title}, {language}, {job_description}, {resume}, {max_questions})

---

🧠 Troubleshooting

• No AI responses: Check your GEMINI_API_KEY
• Mic issues: Allow microphone access in browser
• Parsing failed: Use a text-based (not scanned) PDF