📄 AI Resume Screener using NLP

This is a simple yet powerful AI-driven web application that helps screen resumes against a job description using Natural Language Processing (NLP). It calculates a match score, highlights potential resume improvements, and supports batch upload of multiple PDFs. ▶️ Live App on Streamlit

🔍 Features

• 📤 Upload multiple resumes (PDFs)
• 🧠 NLP-based similarity matching using TF-IDF + Cosine Similarity
• 📌 Compares resumes against a provided Job Description
• 📊 Match score with progress bars and recommendations
• 🔎 Highlights potential missing keywords
• 🎨 Clean Streamlit UI with sidebar and enhancements

---

🚀 How it Works

1. Upload resumes (PDF format)
2. Job description is loaded from a job_description.txt file
3. App cleans and analyzes the text using NLP
4. Calculates a similarity score between each resume and the JD
5. Displays recommendations:
   • ✅ Strong Match (Score ≥ 70%)
   • ⚠️ Needs Improvement (Score < 70%)
6. Highlights missing keywords (basic rule-based logic)

---

🛠️ Tech Stack

• Python
• Streamlit
• Scikit-learn (TF-IDF + Cosine Similarity)
• pdfminer.six (PDF text extraction)

---

📁 Project Structure

resume-screener-nlp/ ├── app.py # Main Streamlit app ├── utils.py # Text extraction, cleaning, similarity functions ├── job_description.txt # Sample JD file to compare resumes ├── requirements.txt # Dependencies └── sample_resumes/ # (Optional) Folder for testing resumes

---

✅ To Run Locally

1. Clone the repository
2. Install dependencies
   pip install -r requirements.txt
3. Add your job description in a job_description.txt file
4. Run the app

streamlit run app.py

---

📌 Sample

Resume File	Match Score	Recommendation
resume1.pdf	82.3%	✅ Strong Match
resume2.pdf	56.7%	⚠️ Needs Improvement

---

💡 Future Ideas

• AI-powered resume feedback (LLM-based)
• PDF/Markdown export of resume scores
• JD upload option
• Skill visualization or radar charts

---

🧑‍💻 Author

Made with ❤️ by Prakash Sharma 

---

📜 License

This project is open-source and free to use under the MIT License.