README.md

metadata

title: ATS Resume Intelligence Engine
emoji: 📄
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860
pinned: false
license: mit
suggested_hardware: cpu-upgrade

ATS Intelligence Engine

An AI-powered Applicant Tracking System that scores resumes against job descriptions using semantic similarity.
Built with FastAPI, LangChain, and Hugging Face — deployable as a Docker Space.

---

Features

• Upload resume & job description (PDF or text)
• Extract text using PyMuPDF
• Semantic similarity via Sentence Transformers
• Keyword matching for skill overlap
• Final ATS score with LangChain + Hugging Face LLM feedback
• REST API built with FastAPI
• Docker Space ready (port 7860)

---

Tech Stack

Layer	Technology
Backend	FastAPI, Uvicorn
ML / Embeddings	Sentence Transformers, Scikit-learn
LLM / Feedback	LangChain, Hugging Face Inference API
PDF Parsing	PyMuPDF
Frontend	HTML / JS (served by FastAPI)
Deployment	Docker, Hugging Face Spaces

---

Deploy on Hugging Face Spaces (Docker)

1. Create a new Space

1. Go to huggingface.co/new-space
2. Choose Docker as the SDK
3. Pick a name (e.g. ats-resume-intelligence)
4. Hardware: CPU upgrade (recommended for Sentence Transformers)

2. Push this repository

git remote add space https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE_NAME
git push space main

Or clone the empty Space repo and copy these files into it, then commit and push.

3. Add your Hugging Face token (required for AI feedback)

In the Space → Settings → Repository secrets, add:

Secret name	Value
HF_TOKEN	Your Hugging Face access token (create one) with Inference permission

The Space will rebuild automatically after you push or add secrets.

4. Open your live app

Your Space URL will be:

https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE_NAME

The UI and API run on the same origin (port 7860), so no CORS or proxy setup is needed.

---

Run locally with Docker

Prerequisites

• Docker
• Hugging Face API token

Setup

Create a .env file:

HF_TOKEN=your_huggingface_token_here

Start

docker compose up --build

Open http://localhost:7860

---

API Endpoints

POST /predict/ats — JSON

{
  "resume_text": "...",
  "job_description": "..."
}

POST /predict/ats/upload — multipart

Field	Type	Description
resume_pdf	File	Resume PDF
resume_text	string	Resume text
jd_pdf	File	Job description PDF
job_description	string	Job description text

Response

{
  "semantic_score": 0.82,
  "keyword_score": 0.74,
  "final_ats_score": 0.79,
  "summary": "..."
}

GET /health — health check

---

Development (without Docker)

pip install -r requirements.txt
set HF_TOKEN=your_token_here
uvicorn main:app --reload --port 7860

For local dev without the Space URL, in Frontend/app.js set:

const API = 'http://localhost:7860';

---

Environment variables

Variable	Description
HF_TOKEN	Hugging Face API token (Space secret or .env)

---

License

MIT License