Overview
A web app that scores a résumé the way an applicant tracking system would, then tells the candidate specifically what to improve.
The problem
Many employers filter résumés through applicant tracking systems before a person reads them.
Candidates are rejected without feedback and cannot tell whether structure, keywords or missing sections are the problem.
The solution
The app parses a pasted or uploaded résumé (.txt or .pdf) into structured signals — skills, education, certifications, contact details, section presence and length.
A scikit-learn model blended with transparent rules produces a score and a list of specific, actionable recommendations. Keywords, weights and thresholds are configurable.
Engineering highlights
Feature extraction pipeline
Raw résumé text is converted into structured, explainable signals before scoring.
ML plus transparent rules
A trained model and readable rules are blended so every score can be explained.
Configurable scoring
Keywords, weights and thresholds can be tuned without retraining.
Containerized deployment
Docker-ready and deployed publicly on Hugging Face Spaces.
What was built
- Résumé parsing for text and PDF
- ML scoring model and rule engine
- Actionable improvement suggestions
- Streamlit web interface
- Docker configuration
- Public deployment on Hugging Face Spaces
Results
- Publicly available as a live demo on Hugging Face Spaces