Resume Screening AI**

**Deploy Link :https://resume-screening-ai-m2duh4xr8i27d9ydvxr7uk.streamlit.app/

Automatic resume ranking system** that helps recruiters quickly shortlist candidates

by comparing resumes against job descriptions using natural language processing.

---

Project Overview

Recruiting is time‑consuming, especially when sifting through hundreds of resumes. This project provides a prototype AI pipeline that:

1. Ingests a corpus of resumes.
2. Cleans and vectorizes text data using industry-standard NLP techniques.
3. Computes similarity between each resume and a target job description.
4. Outputs a ranked list of candidates to streamline hiring decisions.

The implementation is designed as an educational proof‑of‑concept and can be extended to a production‑grade application.

---

Key Features

• Data ingestion and exploratory analysis
• Text preprocessing (lowercasing, punctuation removal, stop‑word filtering)
• TF–IDF vectorization of resume text
• Cosine similarity scoring against a sample job description
• Sorted ranking of candidate resumes

---

Technologies & Dependencies

Category	Tools / Libraries
Language	Python 3.11
Data	pandas, numpy
NLP	NLTK, scikit-learn (TfidfVectorizer)
Environment	Jupyter Notebook / Google Colab

Additional utilities:

• re for regular expressions

A virtual environment is recommended (venv, conda, etc.) to isolate dependencies.

---

Dataset

The dataset used for this project is a simple CSV containing sample resumes mapped to job categories.

• File: data/resume_data.csv (or resume_dataset.csv in earlier notes)
• Records: 169 rows
• Columns: Category, Resume

This repository does not contain private or real candidates; the data is fabricated for demonstration purposes.

---

⚙️ Setup & Installation

1. Clone the repository

   git clone <repo-url>
   cd Resume-Screening-AI

2. Create and activate a virtual environment:

   python -m venv env
   .\env\Scripts\activate   # Windows
   source env/bin/activate     # macOS/Linux

3. Install required packages:

   pip install -r requirements.txt

   If a requirements.txt file is not present, you can manually install:

   pip install pandas nltk scikit-learn

4. Download NLTK resources (if not already available):

   import nltk
   nltk.download('stopwords')

---

📖 Usage

The core logic is implemented in a Jupyter notebook located at: notebooks/Day1_Day2_ResumeScreening.ipynb.

Steps to run:

1. Open the notebook using Jupyter or Colab:

   jupyter notebook notebooks/Day1_Day2_ResumeScreening.ipynb

2. Execute cells sequentially. Key sections of the notebook:

   • Data loading & exploration
   • Text cleaning and preprocessing
   • TF–IDF vectorization
   • Job description matching and ranking

3. Modify the sample job description or extend the notebook with additional features (e.g. PDF parsing, classification models).

---

📊 Output Examples

• TF–IDF matrix (sparse matrix representation)
• Similarity scores for each resume relative to the job description
• Ranked candidates (highest-scoring resumes first)

Screenshots or exported CSVs can be added here for clarity.

---

Architecture & Workflow

System Architecture

flowchart TB
    subgraph Data Layer
        A[Resume CSV Data] --> B[Data Ingestion]
        C[Job Description Input] --> B
    end
    subgraph Processing Layer
        B --> D[Text Preprocessing]
        D --> E[TF-IDF Vectorization]
        E --> F[Similarity Calculation]
    end
    subgraph Output Layer
        F --> G[Ranked Resume List]
    end

    classDef data fill:#f9f,stroke:#333,stroke-width:1px;
    class A,C data;

Loading

Workflow

flowchart LR
    1[Start] --> 2[Load dataset]
    2 --> 3[Clean and preprocess text]
    3 --> 4[Generate TF-IDF features]
    4 --> 5[Specify job description]
    5 --> 6[Compute cosine similarity]
    6 --> 7[Sort resumes by score]
    7 --> 8[End - Output ranked list]

Loading

The diagrams above illustrate the major components and data flow of the project, highlighting how resumes are transformed and compared against job descriptions.

---

Future Enhancements

The current prototype can be expanded along several dimensions:

1. Resume file support – parse PDF/DOCX documents instead of plain text.
2. Classification/Matching model – train supervised algorithms to predict best-fit roles.
3. Web UI – deploy as a Streamlit/Flask/Django application.
4. Explainability – highlight keywords and phrases responsible for scores.
5. Integration with ATS – connect the pipeline to Applicant Tracking Systems.

---

🤝 Contributing

Contributions are welcome! Please follow these guidelines:

1. Fork the repository and create a feature branch.
2. Make changes and include tests where appropriate.
3. Submit a pull request with a clear description of your work.

---

📄 License

This project is licensed under the MIT License.

---

📬 Contact

For questions or feedback, reach out to:

• Author: Kavana S Harthal,Preethi kumari,Tejaswini r

---

Thank you for exploring the Resume Screening AI project! 🙌