This post outlines what it does, how I built it, and how it can help students and job seekers optimize their resumes for better chances of selection.

The Idea

Most candidates submit a generic resume everywhere. But many companies use keyword-matching algorithms or applicant tracking systems (ATS) to filter resumes before a human even sees them. So, I decided to build a smart tool that:

• Uploads your resume as a PDF

• Accepts a job description

• Highlights which keywords are missing

• Calculates a matching success score

All of this — in seconds — via a simple browser app!

Tech Stack

Features

• Upload resume in PDF format

• Paste any job description text

• Compares resume against JD using token-based matching

• Uploads resume securely to an AWS S3 bucket

• Shows success rate based on keyword matching

• Suggests missing keywords to improve the resume

• Simple, clean, and fast user experience

How It Works (End-to-End)

1. Upload & Parse Resume

Users upload their resume in PDF format. Instead of using OCR (which is often inaccurate for digital PDFs and needs system-level dependencies), we extract direct text using libraries like PyMuPDF (fitz) or PDFMiner. This ensures more reliable and faster extraction.

2. Match Against Job Description

We take the pasted JD text, clean and tokenize both the JD and the resume, remove stopwords, and perform intersection matching to calculate:

• Matched keywords

• Success rate (percentage match)

• Missing terms to improve resume

3. Store Resume in AWS S3

We also upload the file to an AWS S3 bucket to simulate real-world enterprise workflows where resumes are stored for future reference or ML training.

File Structure

bashCopyEditresume-analyzer/
│
├── app.py               
├── jd_matcher.py        
├── resumeparser.py      
├── requirements.txt     
├── packages.txt    
├── setup.sh             
├── render.yaml

Key Concepts Used

• Set-based keyword matching

• Stopword removal

• AWS S3 file upload with boto3

• PDF text parsing with PyMuPDF

• Dynamic UI with Streamlit

• Render deployment pipeline

Challenges Faced

• Avoiding system-level dependencies like Tesseract OCR for simplicity

• Cleaning the text for better token matching

• Securely integrating AWS S3 for file uploads

• Ensuring the app works smoothly on cloud (Render)

What I Learned

• Building cloud-hosted tools with a simple UI using Streamlit

• Working with AWS S3 via Python's boto3

• Designing resume matching logic without complex ML models

• Deploying full apps with environment configs on Render

Live Demo

Visit the Live App

GitHub Repo

View Source Code on GitHub

Future Improvements

• Add basic ML-based semantic matching

• Save and download improved resumes

• User authentication

• History tracking of uploaded resumes

Final Thoughts

This project is a great example of how simple technologies — used right — can create powerful tools. If you're a fresher or student looking to improve your resume, try it out! And if you're a developer, feel free to fork and build on top of it.

Thanks for reading!
— Mokka Madan Mohan