📄 Resume Parser Service (GPT-3 Powered)

A GPT-3 based Resume Parser REST API that converts resume PDFs into clean, structured JSON files.
This service accurately extracts key fields such as contact information, education, job experience, and project history.

Example:

• 📄 Input PDF
• 🗂️ Parsed JSON Output

---

🚀 Features

• ✅ Fast and accurate resume parsing using GPT-3 (text-davinci-002)
• ✅ JSON response ready for ATS and HR system integration
• ✅ Supports most common resume formats
• ✅ Flask-powered REST API for easy deployment
• ✅ Environment-based OpenAI key management

---

💰 Token Usage & Cost

Parsing a resume typically takes ~15 seconds and consumes:

• 1500 tokens ($0.03)
• Larger resumes may consume up to 3000 tokens ($0.06)

(Based on OpenAI’s pricing model for the text-davinci-002 engine.)

⚡ Fine-tuning GPT-3 could further improve accuracy, but out-of-the-box performance is already excellent.

---

⚙️ Quick Start

1. Install Prerequisites

• Python 3.7+
• pip3

(For macOS users, see additional instructions below.)

2. Setup

git clone https://github.com/YaroslavShved25/-Resume-Parser-Service-NLP-.git
cd resume-parser-service

# Upgrade pip if needed
python3 -m pip install --upgrade pip

3. Build the Project

./build.sh

4. Set Your OpenAI API Key

• Create a .env file:

OPENAI_API_KEY=YOUR_API_KEY

OR

• Set it as an environment variable:

export OPENAI_API_KEY=YOUR_API_KEY

5. Run the Service

./run.sh

Access the API: http://0.0.0.0:5001/

---

🖥️ macOS Setup Notes

You must install:

• Xcode (run it at least once to complete setup)
• Command Line Tools:

xcode-select --install

• Homebrew
• Python 3 via:

brew install python

---

📂 Supported Fields

🎯 Basic Information

• First Name
• Last Name
• Full Name
• Email
• U.S. Phone Number
• Location
• Portfolio Website URL
• LinkedIn URL
• GitHub URL

🎓 Education

• University
• Degree Level
• Graduation Year & Month
• Majors
• GPA

💼 Job Experience

• Job Title
• Company
• Location
• Duration
• Job Description

🛠️ Project Experience

• Project Name
• Project Description

---

📄 License

This project is licensed under the MIT License.

---

🙋‍♂️ Questions or Contributions?

Feel free to reach out or open an issue. Pull requests are welcome!