rescreener

AI-Powered Resume Screener (Classification + NLP + LLM)

• NLP stands for Natural Language Processing
• LLM stands for Large Language Model

Problem: Automatically classify or rank resumes based on job description fit.

Project Structure

This project follows the following standard project layout structure:

    app_name
    ├── pyproject.toml
    ├── poetry.lock 
    ├── README.md
    ├── .gitignore
    │
    ├── app_name/
    │   ├── __init.py__
    │   ├── main.py
    │   ├── module_a.py
    │   └── package/
    │       ├── __init.py__
    │       └── module_b.py
    │
    └── tests/
        ├── __init__.py
        ├── test_main.py
        ├── test_module_a.py
        │   └── package/
        │       ├── __init.py__
        │       └── test_module_b.py

Prerequisite Tools

• pipx 1.7.1 or later
• poetry 2.1.3 or later
• pylint 3.3.7 or later
• python3 3.13.0 or later
• pytest 8.3.4 or later
• IDE (e.g., PyCharm)

Key Libraries

Stack:

• Data: Sample resumes + job descriptions (or scrape)
• ML: Embedding with BERT + classifier (e.g., logistic regression or fine-tuned transformer)
• Backend: FastAPI to score resumes
• Frontend: Upload interface + match score
• Bonus: Use OpenAI or Hugging Face LLM to generate feedback comments
• Claude 3.5 Sonnet, GPT-4.1

🧩 Project Blueprint: End-to-End ML Flow

Here’s a general pipeline you should follow regardless of the problem domain:

Step	What You Do	Tools to Use
1. Define the problem	E.g., classification, regression, recommendation	N/A
2. CollectData	Load dataset: resume.pdf and jd.txt	Upload resume.pdf and jd.txt
2. Collect/Clean Data	Load, clean, and analyze your dataset	Upload resume.pdf and jd.txt
3. EDA & Feature Engineering	Understand and visualize patterns	Seaborn, Matplotlib
4. Train/Test Split	Prepare train/validation sets	Scikit-learn
5. Build & Train Models	Try several models, tune hyperparameters	Scikit-learn, XGBoost, PyTorch
6. Evaluate Models	Use metrics like accuracy, F1, RMSE, ROC	Scikit-learn
7. Deploy the Model	Create an API to serve predictions	Flask or FastAPI
8. Build UI or App	Web app to interact with the model	Streamlit, React, Dash
9. Monitor & Iterate	Add logging, test edge cases, improve UX	MLflow, logging libs

Input Dataset

In this project we're dealing with unstructured text data — like:

• Candidate resumes (PDFs or raw text)
• Job descriptions (text with skills, roles, requirements)

Setup of Development Environment

Refer to SETUP.md.

ChapGPT

NOTE: This project was created with the help of ChapGPT

Claude Code

• Claude Sonnet 4

npx install -g @anthropic-ai/claude-code

---

Rubens Gomes