Skip to content

Repository files navigation

AI-Based Resume Screening System

An AI-powered recruitment assistant that uses Retrieval-Augmented Generation (RAG) to match candidates with job openings and provide natural-language insights to recruiters.

Overview

This system helps recruiters screen resumes faster by combining local LLM inference with RAG-based retrieval over resume and job data. Recruiters can query candidate information using natural language instead of manually filtering spreadsheets.

Key Features

  • AI-powered resume analysis — suggests suitable job matches for candidates based on skills and profile
  • RAG-based retrieval over resume and job posting data for accurate, context-aware responses
  • Natural language queries — recruiters can ask things like "show me top-ranked candidates" or "who applied most recently"
  • Automated candidate profiling with Firebase integration
  • Fast response time — ~5-8 seconds per query

Tech Stack

  • LLM: Ollama (phi3:mini)
  • Retrieval: RAG (Retrieval-Augmented Generation)
  • Database: Firebase
  • Frontend: Flutter (Dart) — cross-platform (Android, iOS, Web, Windows, macOS)

How It Works

  1. Resumes and job data are ingested and indexed for retrieval
  2. Recruiter submits a natural language query (e.g. "shortlisted candidates for backend role")
  3. RAG pipeline retrieves the most relevant resume/job chunks
  4. Ollama (phi3:mini) generates a natural language response using retrieved context
  5. Results are returned to the recruiter within 5-8 seconds

Use Cases

  • Retrieve top-ranked candidates for a specific role
  • View shortlisted or most recent applicants
  • Get AI-suggested job matches based on candidate skills

Project Context

Developed as a personal project (March 2025 - August 2025) to explore practical applications of LLMs and RAG in real-world recruitment workflows.

About

AI-powered recruitment assistant using RAG and LLM to match candidates with jobs via natural language queries. Built with Flutter and Firebase.

Topics

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages