Smart Resume Screening System with ATS (Advanced Version)

📌 Project Overview:

  This is an advanced-level Smart Resume Screening System powered by ATS (Applicant Tracking System) techniques. It helps HR and recruiters to automatically screen, score, and shortlist resumes based on a given Job Description (JD). It also includes a Flask web interface, stores data in a SQLite database, and uses NLP and ML techniques like Named Entity Recognition (NER), TF-IDF, Cosine Similarity, and Job Role Classification using spaCy and other ML models.

---

📁 Project Structure:

resume_screening_system/
├── app/
│   ├── main.py               # Main workflow script
│   ├── parser.py             # Extracts resume text
│   ├── ats_scorer.py         # Scores resumes based on job description
│   ├── mailer.py             # Sends emails to shortlisted candidates
│   ├── file_manager.py       # Handles file movements and storage
│   ├── ner_extractor.py      # Extracts name, email, skills using spaCy NER
│   ├── role_classifier.py    # Predicts job role category using ML model
│   ├── database.py           # SQLite database operations (insert, query, create)
│   └── flask_app.py          # Flask UI for uploading and viewing results
├── resumes/                  # Raw resumes input folder
├── shortlisted/              # Shortlisted resumes output folder
├── data/
│   ├── job_description.txt   # Job Description for screening
│   ├── ranked_candidates.csv # Output: ATS score and details
├── utils/
│   ├── text_cleaner.py       # Text preprocessing
│   ├── similarity.py         # TF-IDF and cosine similarity
├── logs/
│   └── ats_log.log           # Logs for system events
├── config/
│   └── email_config.json     # SMTP configuration for email
├── templates/
│   └── index.html            # HTML template for Flask app
├── static/                   # CSS/JS if needed for Flask UI
├── test/                     # Unit test files
├── requirements.txt          # List of required packages
└── README.md                 # Project documentation

---

⚙️ Workflow (Advanced Pipeline):

1. User uploads resumes using the Flask UI.
2. parser.py reads and extracts raw text from resumes.
3. text_cleaner.py preprocesses the text.
4. job_description.txt is also cleaned similarly.
5. ner_extractor.py extracts entities like name, email, skills using spaCy NER.
6. role_classifier.py predicts the job role category for each resume.
7. ats_scorer.py calculates ATS score using TF-IDF + Cosine Similarity.
8. ranked_candidates.csv stores all info: name, score, predicted role, etc.
9. database.py stores this data in a SQLite database.
10. file_manager.py moves top-scoring resumes to /shortlisted.
11. mailer.py sends emails to shortlisted candidates.
12. Logs are written to ats_log.log.

---

🌐 Flask UI:

• Upload resumes (PDF/DOCX)
• See real-time ATS score, extracted info, predicted role
• Search/filter resumes by score or role

---

🧠 ML & NLP Features

NER Extraction using spaCy for:

• Names
• Emails
• Phone Numbers
• Skills

ATS Scoring using:

• TfidfVectorizer
• cosine_similarity

Role Classification:

• Trained classifier (e.g., Logistic Regression or RandomForest)
• Input: Resume text → Output: Predicted job role

---

🧰 Tools & Libraries

Feature	Libraries/Tools Used
Resume Parsing	pdfminer.six, python-docx, textract
Text Cleaning	re, nltk, spacy, string
NER	spaCy
Similarity Matching	scikit-learn, TfidfVectorizer
Classification	sklearn models, joblib
Database	sqlite3, pandas
Web App	Flask, HTML, CSS, Jinja2
Emailing	smtplib, email.message, ssl
Logging	Python logging module
Testing	unittest, pytest

---

🔋 Setup Instructions

1. Clone the repository

git clone https://github.com/yourusername/resume_screening_system.git
cd resume_screening_system

2. Install requirements

pip install -r requirements.txt

3. Download NLTK and spaCy models

import nltk
nltk.download('punkt')
nltk.download('stopwords')

python -m spacy download en_core_web_sm

4. Run Flask Web App

python app/flask_app.py

---

🧪 Testing:

• Place test files in /test/
• Run with:

pytest test/

---

✅ Future Improvements:

• Use MongoDB for large-scale resume data
• Integrate OpenAI embeddings for semantic similarity
• Add user authentication for HR dashboard
• Resume feedback system for candidates

---

📬 Contact:

Developer: Mahboob Alam Email: [email protected] LinkedIn: https://www.linkedin.com/in/mahboob-alam-242342152

---

⭐ Acknowledgments:

Thanks to the open-source libraries and online resources that helped shape this project!