Staff-plus GenAI Interviews

The Staff-plus GenAI Interviews project is designed to manage and process interview content for Back Market. This project includes functionalities to read, concatenate, and summarize interview content using a large language model in GCP.

🗂️ Project Structure

• interviews/refinement/: Directory containing interview audio transcriptions in text files (.txt).
• interviews/questions/: Directory containing questions in text files (.txt).
• interviews/audio/: Directory containing audio files (.mp3).
• prompts/: Directory containing prompt templates for the language model (.yaml).
• out/: Directory where the output files are saved (.md).
• pyproject.toml: Project configuration file.
• README.md: Project documentation.

🛠️ Installation

1. Clone the repository:

   git clone https://github.com/backmarket-oss/staff-plus-genai-interviews.git
   cd staff-plus-genai-interviews

2. Install dependencies:

   make install

🚀 Usage

1. Configure Environment Variables: Set the following environment variables:

   • GOOGLE_CLOUD_PROJECT: Google Cloud Project Name
   • DEFAULT_GCP_REGION: Google Cloud Location Name (default: europe-west2)
   • GENAI_MODEL: Model name for the language model (default: google/gemini-1.5-pro-002)
   • GENAI_TEMPERATURE: Temperature setting for the language model (default: 1.0)
   • GENAI_MAX_TOKENS: Maximum tokens for the language model (default: 8192)

2. Run the main script:

   make run

🧪Methodology

1. Interview selection

2. Interview audio AI processing

3. Interview text review

4. Interview text summary

Results

[extract_and_summarize_interviews_content_back_market.md]

📜 License

This project is licensed under the Apache License 2.0. See the LICENSE file for more details.