ResumeCraft-AI

• 

  cover

🚀 Inspiration

Creating a professional personal website is still hard for many people — especially students, job seekers, and non-technical professionals. Most already have a résumé, but turning that résumé into a modern, visually appealing web profile requires design skills, frontend development, and time.

We asked a simple question:

What if your résumé could instantly become a beautiful personal website?

That idea became ResumeCraft-AI.

💡 What it does

ResumeCraft-AI takes a PDF résumé as input and automatically generates a colorful, modern, and responsive personal website that reflects everything available in the CV — including:

Name & contact information

Education

Skills

Experience

Structured sections ready for the web

The output is a deployable static website (HTML + CSS) that can be hosted instantly using GitHub Pages.

🛠️ How we built it

The project follows a clean AI-powered pipeline:

PDF Resume Upload

OCR Processing

We use PaddleOCR to extract text from résumé PDFs (works even with scanned documents).

Resume Parsing

Extracts structured information (name, email, phone, education, skills, experience).

AI Enhancement (ERNIE via Novita API)

Improves wording, section clarity, and presentation.

Website Generation

Converts structured content into:

Markdown

HTML

Stylish CSS

One-click Deployment

Hosted automatically on GitHub Pages.

🤖 Why ERNIE + PaddleOCR?

This hackathon required real AI usage, not templates.

PaddleOCR ensures accurate text extraction from real-world résumés (including scanned PDFs).

ERNIE (via Novita API) is used to:

Enhance résumé content

Improve readability and structure

Prepare web-ready descriptions

Together, they make ResumeCraft-AI robust, intelligent, and production-ready.

⚙️ Built With

Python

PaddleOCR

ERNIE (Novita API)

pdf2image + Poppler

HTML5

CSS3

GitHub Pages

GitHub Actions

🚧 Challenges we faced

Handling OCR accuracy across different résumé formats

Fixing Poppler & OCR setup on Windows

Adapting to PaddleOCR API deprecations

Parsing unstructured OCR text into meaningful sections

Ensuring the site works as a pure static deployment

Each challenge helped us deeply understand AI pipelines + deployment workflows.

📚 What we learned

End-to-end AI product building

OCR + NLP integration

Static site generation

GitHub Pages CI/CD

Real-world debugging under hackathon pressure

🌍 Impact & Future Scope

ResumeCraft-AI can be extended to:

Multiple website themes

Portfolio sections

LinkedIn/GitHub auto-integration

Multilingual résumé support

Custom domain publishing

Built With

• css3
• ernie
• githubaction
• githubpages
• html5
• novita
• paddleocr
• pdf2image
• poppler
• python