README.md

Legal Considerations of Using LLM on Personal Data

Before implementing AI-driven CV analysis on a productive environment, we advise consulting with a legal specialist to ensure compliance with relevant data protection laws. Key considerations include:

1. Consent and Purpose

   • Obtain clear, explicit consent from individuals before processing their data.
   • Limit data usage strictly to the recruitment process, avoiding any secondary purposes.

2. Data Anonymization

   • Remove or mask personally identifiable information (PII) before processing CVs to minimize privacy risks.
   • This includes names, contact details, and other sensitive identifiers.

3. Review API Policies

   • Familiarize yourself with OpenAI's usage policies to avoid transmitting unnecessary sensitive information.
   • Ensure secure transmission of data to and from the API.

4. Bias and Fairness

   • Regularly review AI outputs for potential bias or unfair patterns in candidate evaluations.
   • Provide additional checks to ensure all candidates are assessed equitably.

5. Security and Transparency

   • Use encryption to protect data during storage and transmission.
   • Clearly inform candidates about how AI is used in the recruitment process and provide options to withdraw their data if needed.

Project Overview

The project consists of a Jupyter notebook that:

• Extract text from PDF files containing candidate CVs and a job position description.
• Normalize and structure the extracted text using OpenAI's GPT model.
• Evaluate candidates by matching their CVs against the job position description.
• Output the results in JSON format for analysis in Power BI.

Prerequisites

Before you begin, ensure you have the following:

1. Installing Python 3.7 or Higher: Download Python
2. Installing Visual Studio Code (VS Code) a. VS Code Official Download Page: Download Visual Studio Code b. Microsoft: Set up VS Code - Official setup overview from Microsoft.
3. Installing Git a. Git for Windows: Download Git

Setting up the project

1. Cloning the GitHub Repository

If you're unfamiliar with Git, these steps will guide you through cloning the repository using VS Code.

1. Open VS Code.
2. Open the Command Palette by pressing Ctrl+Shift+P.
3. Type "Git: Clone" and select it.
4. Enter the Repository URL: https://github.com/OscarValerock/BIBB-PBI-CV-AI-Analysis.git
5. Choose a Local Directory: Select a folder on your computer to store the project.
6. Open the Repository: VS Code will prompt you to open the repository once cloned. Click Open.

2. Setting Up a Virtual Environment in VS Code

Once you've cloned the repository and have your project open in Visual Studio Code, it's best practice to create a virtual environment for your project. This isolates the required Python packages, making your setup more stable and organized. Here's how to do it using the VS Code Command Palette:

• In the Command Palette (Ctrl+Shift+P), type Python: Create Environment and select Create Environment.
• VS Code will prompt you to select a folder. Select your project folder (the cloned repository) and choose venv as the virtual environment type.
• Accept to install the packages from requirements.txt

3. Add the Constants.py File

Create a file named Constants.py in your project's root directory. This file will store your OpenAI API key.

OpenAIKey = "your-openai-api-key"

Important: Replace "your-openai-api-key" with your actual OpenAI API key. Keep this file secure and avoid sharing it publicly.

graph TD
    subgraph "VS Code - Jupyter"
        A[Start]
        B[Extract text from CV PDFs]
        A --> B
        B --> C[OCR_Results.json]
        F[Extract text from Position PDF]
        A --> F
        F --> G[OCR_Position.json]
        
        E[LLM_Normalized_CV.json]
        I[LLM_Position.json]
        K[LLM_Analysis.json]
        
        subgraph "OpenAI"
            D[Summarize CVs]
            H[Normalize Position Description]
            J[Evaluate Candidates]
        end
        
        C --> D
        D --> E
        G --> H
        H --> I
        E --> J
        I --> J
        J --> K
    end
    
    E --> L[Analyze with Power BI]
    I --> L
    K --> L
    
    %% Define styles
    classDef openai color:#FFFFFF, fill:#FF0000,stroke:#FF0000,stroke-width:2px;
    classDef vscode color:#FFFFFF, fill:#008000,stroke:#008000,stroke-width:2px;
    classDef powerbi color:#000000, fill:#FFD700,stroke:#FFD700,stroke-width:2px;

    %% Apply styles
    class D,H,J openai
    class A,B,C,E,F,G,I,K vscode
    class L powerbi

    

Loading