JobIntel Backend

Overview

JobIntel Backend is a robust, scalable RESTful API built with Flask that powers job-resume matching and advanced skill extraction. It leverages NLP techniques such as TF-IDF vectorization, sentence embeddings (via Sentence Transformers), and spaCy/NLTK-based skill extraction to semantically analyze resumes and job descriptions for effective talent-job alignment.

---

Features

• TF-IDF Based Resume-to-Job Matching
• Semantic Matching with Sentence Embeddings
• Multi-Method Skill Extraction
  • Regex-based
  • spaCy-based NLP
  • Semantic similarity-based
• Skill categorization into predefined domains (e.g., Programming, Cloud, Web Dev, etc.)
• Memory-aware model loading for efficient operation on low-resource machines

---

Tech Stack

• Backend: Flask, Flask-CORS
• NLP Libraries: spaCy, NLTK, Sentence Transformers
• ML: scikit-learn, TF-IDF, Cosine Similarity
• Data: JSON-based Job Dataset
• Utilities: psutil, re, gc, pandas, numpy

---

Directory Structure

.
├── app/                    # Core Flask app directory
│   ├── job_matcher.py      # JobMatcher class (TF-IDF + Semantic)
│   ├── skill_extractor.py  # SkillExtractor class
│   └── routes.py           # API routes (if applicable)
├── data/
│   └── jobs_descriptions.json  # Job dataset
├── main.py                 # Entry point for the Flask app
├── requirements.txt        # Dependency list
└── README.md               # This file

---

Setup Instructions

1. Clone the Repository

git clone https://github.com/yourusername/jobintel-backend.git
cd jobintel-backend

2. Create a Virtual Environment

python -m venv venv
source venv/bin/activate  # For Windows: venv\Scripts\activate

3. Install Dependencies

pip install -r requirements.txt

If spaCy throws a model loading error, download the model manually:

python -m spacy download en_core_web_sm

4. Run the Flask App

python main.py

---

API Endpoints

Base URL: http://localhost:5000/

Endpoint	Method	Description
/match	POST	Match resume text to jobs (TF-IDF/Semantic)
/extract-skills	POST	Extract and categorize skills from resume text
/health	GET	Health check for backend status

---

JSON Input Format

/match

{
  "resume_text": "Skilled in Python, Django, and REST APIs...",
  "method": "tfidf" // or "semantic"
}

/extract-skills

{
  "text": "Experienced in AWS, Docker, and Kubernetes."
}

---

Output Format

/match Response

[
  {
    "job_id": "101",
    "title": "Machine Learning Engineer",
    "score": 0.8723,
    "description": "Looking for an ML engineer with Python and TensorFlow..."
  }
]

/extract-skills Response

{
  "all_skills": ["aws", "docker", "kubernetes"],
  "categorized_skills": {
    "cloud_devops": ["aws", "docker", "kubernetes"],
    "programming": []
  },
  "skill_count": 3,
  "extraction_methods": {
    "regex": 3,
    "spacy": 2,
    "semantic": 1
  }
}

---

Dataset Format

File: data/jobs_descriptions.json

[
  {
    "id": "101",
    "title": "Data Scientist",
    "description": "We are looking for a data scientist skilled in Python, Pandas..."
  }
]

---

Performance Considerations

• Uses conditional model loading (spaCy/SentenceTransformer) based on available RAM
• TF-IDF vectorization is memory-optimized using max_features=500
• Garbage collection (gc) is manually triggered for memory cleanup

---

Deployment (Optional)

To run in production:

gunicorn main:app --bind 0.0.0.0:5000 --workers 4

---

License

This project is licensed under the MIT License. See the LICENSE file for details.

---