An intelligent, Streamlit-based application that rewrites and optimizes resumes to better align with job descriptions using GPT-4. It provides keyword match scores, skill alignment analysis, and generates a professionally tailored PDF resume.
✅ Upload Resume and Job Description in .pdf, .doc, .docx, or .txt formats
✅ Intelligent tailoring of resume content using OpenAI GPT-4
✅ Retains important technical keywords even if they are not present in the JD
✅ Match score analysis based on keyword similarity
✅ Soft skills and technical skills alignment scoring
✅ Actual page count from generated PDF resume
✅ Download tailored resume as a polished PDF
✅ Resume history stored in a local SQLite database
✅ Clean and responsive Streamlit UI
AI-Powered Resume Assistant/
│
├── resume.py # Main Streamlit application
├── .env # OpenAI API Key (excluded from Git)
├── requirements.txt # Python dependencies
├── README.md # Project documentation
├── JD.jpeg # Screenshot: JD Upload
├── Metrics.jpeg # Screenshot: Metrics Output
└── resume_history.db # Local SQLite DB for resume storage (auto-created)
git clone https://github.com/<username>/AI-Powered-Resume-Assistant.git
cd AI-Powered-Resume-Assistant
pip install -r requirements.txtCreate a .env file in the root folder with:
OPENAI_API_KEY=your_openai_api_key_herestreamlit run resume.py- Python 3.8+
- LibreOffice (optional, for
.docfile conversion)
- Resume keyword cloud visualization
- JD keyword extraction summary
- Option to fine-tune optimization aggressiveness
- Multi-language support
This project is intended for learning and personal use only. Commercial distribution is not permitted.
Built to address the challenge of customizing resumes for today's competitive job market.
Generative AI • Agentic AI • Azure PaaS • Cloud-Native .NET Solutions
This project was developed as a practical Proof of Concept demonstrating:
- Multi-Agent AI
- LangGraph Orchestration
- Retrieval-Augmented Generation (RAG)
- Memory Management
- Guardrails & Observability
- Production-Inspired Architecture Patterns

