Talent Talk

An AI-powered interview system that conducts technical interviews, evaluates responses, and generates comprehensive reports.

Revolutionizing technical interviews with AI-driven conversations and insights

Features

• Multiple Interview Modes: Text-based or voice-based interviews
• Dynamic Resume Analysis: Upload any resume for personalized questions
• Customizable Questions: Upload your own interview questions
• Voice Interaction: Speak with the AI interviewer using Whisper (local) or AssemblyAI (cloud)
• Automatic Evaluation: Get detailed scoring and feedback on interview performance
• Report Generation: Comprehensive HR reports with candidate assessment
• PDF Export: Download professional PDF reports

Components

• LangGraph Workflow: Multi-stage interview process with evaluation and reporting
• Speech-to-Text: Local (Whisper) or cloud-based (AssemblyAI) options
• Text-to-Speech: Voice responses using ElevenLabs
• Vector Search: Dynamic resume and question analysis
• PDF Generation: Professional report formatting and export

Installation

1. Clone the repository:

git clone https://github.com/yourusername/talent-talk.git
cd talent-talk

2. Install dependencies:

pip install -r requirements.txt

3. Set up environment variables:

cp .env.example .env
# Edit .env with your API keys

Usage

Text-based Interview

streamlit run text_interview_app.py

Voice-based Interview with Whisper (Local)

streamlit run voice_interview_app_whisper.py

Voice-based Interview with AssemblyAI (Cloud)

streamlit run voice_interview_app_assemblyai.py

Interview Process

1. Setup: Configure interview parameters and upload resume/questions
2. Introduction: AI introduces itself and asks candidate to introduce themselves
3. Resume Questions: AI asks about specific projects from the candidate's resume
4. Technical Questions: AI asks technical questions related to the position
5. Follow-up Questions: AI may ask follow-up questions for incomplete answers
6. Evaluation: After the interview, AI evaluates the candidate's responses
7. Report Generation: AI creates a comprehensive HR report
8. PDF Export: Download the report as a professional PDF

Configuration Options

• Interviewer Mode: friendly, formal, or technical
• Position: Job title for the interview
• Company Name: Company conducting the interview
• Number of Questions: How many technical questions to ask
• Follow-up Questions: How many follow-up questions to ask
• Voice Settings: Model size and language options (for Whisper)

API Keys

This project uses the following external APIs:

• AssemblyAI: For cloud-based speech recognition
• ElevenLabs: For realistic AI voice responses
• Google Generative AI: For the LLM powering the interview

Directory Structure

talent-talk/
├── src/                    # Core source code
│   ├── workflow.py         # Basic workflow implementation
│   ├── dynamic_workflow.py # Enhanced workflow with dynamic resume
│   └── pdf_utils.py        # PDF generation utilities
├── utils/                  # Utility functions
│   └── audio_utils.py      # Audio processing utilities
├── data/                   # Default data files
│   ├── default_questions.pdf  # Default interview questions
│   └── default_resume.pdf     # Default resume for testing
├── generated_reports/      # Generated PDF reports
├── uploaded_resumes/       # User-uploaded resumes
├── uploaded_questions/     # User-uploaded questions
├── text_interview_app.py   # Text-based interview app
├── voice_interview_app_whisper.py    # Voice app with Whisper
├── voice_interview_app_assemblyai.py # Voice app with AssemblyAI
├── requirements.txt        # Project dependencies
├── .env.example            # Example environment variables
├── .gitignore              # Git ignore file
└── README.md               # Project documentation

Requirements

• Python 3.8+
• Streamlit
• LangGraph
• LangChain
• Google Generative AI
• Whisper (for local voice processing)
• AssemblyAI (for cloud voice processing)
• ElevenLabs (for voice responses)
• FPDF (for PDF generation)

License

MIT

Acknowledgments

• OpenAI for Whisper
• AssemblyAI for speech recognition
• ElevenLabs for text-to-speech
• LangChain and LangGraph for the agent framework