surendrallam/deltagenai_takehome

A solution repo for Deltagen's AI take home

Personalized Email Generation and Comparison

This project generates personalized emails for leads using OpenAI's GPT-3.5 models and compares the generated emails to determine the best one to send. The results are saved in a JSON file.

Approach

1. Load API Key: The OpenAI API key is loaded from an environment variable.
2. Load CSV File: The CSV file containing lead information is loaded into a pandas DataFrame.
3. Extract Relevant Details: Relevant details such as Name, Email, Company, Job Title, and Industry are extracted from the DataFrame.
4. Generate Emails: Personalized emails are generated for each lead using two different GPT-3.5 models.
5. Compare Emails: The generated emails are compared based on relevance, tone, and engagement potential to determine the better email.
6. Save Results: The results, including the generated emails and the selected email, are saved in a JSON file.

How to Run

1. Set Up Environment:

   • Ensure you have Python installed.
   • Install the required packages using the following command:

     pip install pandas openai

2. Set OpenAI API Key:

   • Set your OpenAI API key as an environment variable:

     export OPENAI_API_KEY='your-api-key'

3. Prepare CSV File:

   • Place your CSV file containing lead information in the ./data/ directory. The CSV file should have columns: Name, Email, Company, Job Title, and Industry.

4. Run the Script:

   • Execute the script to generate and compare emails:

     python personalised_mail_generation.py

5. Check Results:

   • The results will be saved in a JSON file located at ./data/personalized_emails.json.

Example CSV File

Name,Email,Company,Job Title,Industry
John Doe,john.doe@example.com,Example Inc.,CEO,Technology
Jane Smith,jane.smith@example.com,Sample Corp.,CTO,Finance
...