Projects/AI Resume Screening System
AI / MLFinal Year

AI Resume Screening System

NLP-based resume ranking engine with a recruiter dashboard.

Case Study

Problem Statement

HR teams and campus placement cells receive hundreds of resumes per opening and spend hours manually shortlisting candidates, which delays hiring and often misses well-qualified applicants buried in the pile.

Our Solution

This project parses resumes (PDF/DOCX), extracts skills, education, and experience using NLP, then scores each resume against a job description using a similarity model. Recruiters get a ranked dashboard instead of a raw folder of files.

Features

  • Resume parsing for PDF and DOCX formats
  • Skill extraction using spaCy NER
  • Job-description-to-resume similarity scoring (TF-IDF + cosine similarity)
  • Recruiter dashboard with ranked candidate list and filters
  • Exportable shortlist as CSV

Technology Stack

PythonspaCy / NLTKscikit-learnFlaskReactSQLite

How It's Built

  1. 1

    Upload service accepts resumes and stores raw files

  2. 2

    Parsing service extracts structured text (name, skills, experience)

  3. 3

    Scoring engine compares candidate profile vectors to the job description vector

  4. 4

    Dashboard (React) queries the Flask API and renders ranked results

FAQ

Can I customize the scoring logic?

Yes — the mentorship call covers how to swap TF-IDF for embeddings (e.g. Sentence-BERT) if you want to extend the project.

Does it need a GPU?

No, it runs comfortably on a normal laptop CPU.

Is this project unique or a copy of something online?

It's built specifically for this kit with original code and structure — you're expected to understand and adapt it, not submit it unread.