ATS Resume Scorer

A web app that scores how well a resume matches a job description and returns actionable feedback. Built with FastAPI + Streamlit, using spaCy and Sentence Transformers for NLP and the Groq API for LLM-generated suggestions.

What it does

1. Upload a resume (PDF / DOC / DOCX) and paste a job description.
2. The backend parses the resume, extracts skills and experience, and compares them to the JD using semantic similarity.
3. You get an ATS score, a breakdown by category (formatting, keywords, content, skill validation, ATS compatibility), and LLM-written suggestions for what to improve.
4. Past analyses are saved to your account so you can revisit them.

Tech stack

• Frontend: Streamlit
• Backend: FastAPI (Python)
• NLP: spaCy (en_core_web_md), Sentence Transformers (all-MiniLM-L6-v2)
• LLM: Groq API (Llama 3)
• Auth + Database: Supabase (email/password and Google OAuth)
• PDF report export: WeasyPrint + Jinja2

Project structure

ATS_SCORER/
├── backend/              FastAPI app, NLP services, API routes
├── frontend/             Streamlit app, views, components
├── jupyter notebooks/    Research and dataset prep (not used at runtime)
├── ml model/             Exported ML artifacts
├── requirements.txt      Combined backend + frontend dependencies
└── .env.example          Template for environment variables

Setup

1. Clone and create a virtual environment

git clone <repo-url>
cd ATS_SCORER
python -m venv venv
source venv/bin/activate         # Windows: venv\Scripts\activate

2. Install dependencies

pip install -r requirements.txt
python -m spacy download en_core_web_md

WeasyPrint needs system libraries on Linux:

# Fedora
sudo dnf install -y cairo pango gdk-pixbuf2 libffi

# Debian / Ubuntu
sudo apt install -y libcairo2 libpango-1.0-0 libpangoft2-1.0-0 libffi-dev

3. Configure environment variables

Copy the template and fill in your keys:

cp .env.example .env

You need:

• A Supabase project — grab SUPABASE_URL, SUPABASE_KEY (service role), and SUPABASE_ANON_KEY from Project Settings → API.
• A Groq API key from console.groq.com.
• (Optional) Google OAuth set up in the Supabase dashboard if you want Google sign-in.

The Streamlit frontend also reads Supabase config from frontend/.streamlit/secrets.toml. Copy secrets.toml.example to secrets.toml and fill it in.

4. Run the backend

From the project root:

uvicorn backend.main:app --reload --host 0.0.0.0 --port 8000

The API is now at http://localhost:8000.

5. Run the frontend

In a new terminal (with the venv activated):

streamlit run frontend/streamlit_app.py

The app opens at http://localhost:8501.