ResumeMate 🧠

AI-Assisted Resume Analyzer using Gemini LLMs

ResumeMate is an AI-assisted resume analysis system that compares a candidate’s resume with a job description and generates explainable, ATS-inspired feedback.
It is designed as a decision-support tool, simulating how recruiters and Applicant Tracking Systems (ATS) evaluate resumes — without fabricating experience or hiding logic.

---

🚀 Features

• 📄 Resume and job description analysis (plain text)
• 🎯 ATS-inspired skill matching (heuristic-based)
• 📊 Explainable match score (0–100)
• ✍️ Weak resume bullet detection and rewriting
• 📉 Skill gap identification (critical vs secondary)
• 🤖 ATS keyword improvement suggestions (no stuffing)
• 🧭 Career guidance recommendations
• 🛡️ Robust Gemini API integration with retries, fallbacks, and mock mode

---

🛠 Tech Stack

• Language: Python
• LLM: Google Gemini (gemini-1.5-flash / gemini-2.5-flash)
• Runtime: Kaggle Notebook / Jupyter
• Secrets Management: Kaggle Secrets or environment variables

---

🧠 How ResumeMate Works

1. The user provides:
   • Resume text
   • Job description text
2. The system sends both inputs to Gemini under strict constraints
3. Gemini performs structured analysis:
   • Skill extraction
   • Requirement classification
   • Match scoring with explanation
   • Resume bullet rewriting
4. ResumeMate parses the LLM output using defensive JSON extraction
5. A consolidated, structured report is generated

⚠️ ResumeMate focuses on explainability and reliability, not blind automation.

---

⚙️ How to Run

Option 1: Run on Kaggle (Recommended)

1. Upload resumemate_gemini.ipynb to Kaggle
2. Add your Gemini API key:
   • Go to Account → Secrets
   • Name: GEMINI_API_KEY
3. Enable Allow access to secrets in notebook settings
4. Run all cells

---

Option 2: Run Locally (Optional)

export GEMINI_API_KEY=your_api_key_here

Then open the notebook and run all cells.

📄 Example Output

{
  "match_score": 50,
  "missing_critical_skills": ["REST APIs"],
  "rewritten_bullets":
 [
    {
      "before": "Worked on Python scripts for automation.",
      "after": "Developed Python scripts to automate routine tasks."
    }
  ]
}

🔐 Security & Reliability

API keys are never hardcoded

Supports:

Environment variables

Kaggle Secrets

Includes:

Retry logic with exponential backoff

Mock mode if API key is missing

Hardened JSON parsing to handle LLM inconsistencies

⚠️ Limitations (Important)

ResumeMate does NOT:

Act as a real ATS system

Automatically apply to jobs

Scrape job boards

Fabricate experience or metrics

Replace human judgment

It is intended as a decision-support and learning tool, not a magic solution.

📌 Motivation

This project was built to explore:

Robust LLM system design

Prompt constraint engineering

Explainable AI outputs

Secure and production-aware API integration

Rather than focusing on flashy UI, ResumeMate prioritizes engineering correctness and transparency.

📜 License

This project is released for educational and portfolio use.

👤 Author

J Hope Kumar 📧 kumarhope2018@gmail.com

🔗 LinkedIn: https://linkedin.com/in/12HOPE