How rehanfarishta03/resume-screening-nlp works
This repository provides a Streamlit resume-screening app: a user supplies a resume and job description, the app extracts file text, then computes a lexical match score, keyword gaps, and a predicted job category. A separate training script builds the TF-IDF classifier artifacts consumed by scoring. The README also describes broader notebook-based model comparisons, evaluation, and BERT features; those capabilities are documented but their implementation was not sampled, so they are shown as a separate, unverified branch.
Main entry points: app.py Streamlit interface · train_model.py TF-IDF classifier training · resume Screening using NLP.ipynb Model experiments and evaluation · scoring.py ATS scoring and classification
Components 9
Screening app
- Streamlit interface
app.py
Text analysis
- File text extraction
parsing.py - ATS scoring and classification
scoring.py - Trained model artifacts
artifacts
Model training
- Resume-job dataset
Resume Screening with NLP (TF-IDF)/resume_job_matching_dataset.csv - TF-IDF classifier training
train_model.py
Notebook capabilities
- Model experiments and evaluation
Resume Screening with NLP (TF-IDF)/resume Screening using NLP.ipynb - BERT embeddings
Resume Screening with NLP (TF-IDF)/resume Screening using NLP.ipynb
Other components
- Resume submitter
Connections 8
- Resume submitter → Streamlit interface: submits inputs
app.py - Streamlit interface → File text extraction: extracts uploaded text
app.py - Streamlit interface → ATS scoring and classification: requests analysis
app.py - ATS scoring and classification → Trained model artifacts: loads vectorizer and classifier
scoring.py - TF-IDF classifier training → Resume-job dataset: reads
train_model.py - TF-IDF classifier training → Trained model artifacts: saves trained artifacts
train_model.py - Model experiments and evaluation → BERT embeddings: uses embeddings
README.md - Model experiments and evaluation → TF-IDF classifier training: compares model approaches
README.md