jackboi / tinyllama-resume

Resume Classification Model

Overview
This repository contains a proof-of-concept model for fine-tuning a ‘light’ Large Language Model (LLM) on a resume-to-labor class dataset. The model is designed to classify resumes into various professional categories based on the skills and experiences detailed in the resume.

Model Details
Type: Fine-tuned ‘light’ LLM
Purpose: Resume classification
Training Data: Approximately 4,000 synthetic resumes

View the full labor_categories.json

View a sample of the training data

Labor Categories
The model is trained to classify resumes into the following labor categories:

Business Administration
Contracts Administration
Cyber Security
Cyber Security Technical Analysis
Data Analysis
Data Science
Engineering (General)
Executives
Financial Analysis
Intelligence Analysis

Training Data
The training dataset consists of synthetic resumes created to represent varying skill levels within each of the above labor categories. These synthetic resumes were generated to capture the diverse range of experiences, skills, and qualifications typical of professionals in these fields.

Use Case
This model can be used to:
Automatically categorize job applications
Assist in HR processes for initial resume screening
Help job seekers understand which labor category their resume might fall into

Limitations
As a proof of concept, this model may not capture all nuances of real-world resumes

Performance may vary on resumes from fields not represented in the training data

The model’s accuracy on real-world data should be thoroughly evaluated before any production use

Future Work
Expand the training dataset with more diverse and real-world resumes
Fine-tune the model on additional labor categories
Evaluate and improve the model’s performance on edge cases and multi-category resumes