skill-sync

SkillSync - AI Resume Screener

SkillSync is a Flask-based resume screening app that helps recruiters evaluate candidates faster using AI-assisted scoring, skill matching, and interview-question generation.

Features

• Upload and screen resumes in PDF, TXT, and CSV formats
• AI-based candidate scoring with cosine similarity + model scoring
• Optional keyword boosting for priority skills
• Missing-skill detection and matched-skill highlights
• Auto-generated interview questions with Groq API and local fallback
• Candidate ranking, shortlist/reject workflow, and CSV export
• Interactive dashboard pages with charts and reporting

Project Structure

skillsync/
├── app.py
├── scorer.py
├── parser.py
├── tfidf.py
├── model.py
├── questions.py
├── requirements.txt
├── .env (local only, not committed)
├── templates/
│   ├── index.html
│   ├── screen.html
│   └── report.html
└── static/
    ├── style.css
    └── script.js

Tech Stack

• Python, Flask, Flask-CORS
• PyPDF2, NumPy
• Requests, python-dotenv
• Vanilla JavaScript, Chart.js, THREE.js

Getting Started

1) Clone the repository

git clone https://github.com/Subee2801/Smart-resume-screener.git
cd Smart-resume-screener

2) Create and activate a virtual environment

python -m venv .venv

Windows (PowerShell):

.venv\Scripts\Activate.ps1

macOS/Linux:

source .venv/bin/activate

3) Install dependencies

pip install -r requirements.txt

4) Configure environment variables

Create/update .env:

GROQ_API_KEY=your_groq_api_key_here

.env is ignored by git for security.

5) Run the app

python app.py

Open in browser:

http://127.0.0.1:5000

CSV Input Format

SkillSync supports CSV bulk candidate ingestion. Expected columns are flexible, with best results when using:

• Name
• Email
• Phone
• Skills
• Experience

Each CSV row is treated as one candidate profile and scored like a resume.

API Endpoints

• GET / - Home page
• GET /screen - Screening page
• GET /report - Reporting dashboard
• POST /analyze - Analyze and rank uploaded candidates
• POST /reanalyze - Re-rank with keyword boost weighting
• POST /export - Export shortlisted candidates as CSV

Notes

• Keep API keys only in .env
• Upload files are stored locally in uploads/
• For production use, run behind a WSGI server and add proper auth/logging