Job Matching AI Model

A TensorFlow Lite model for job matching in Flutter apps. Analyzes job descriptions and matches them with user skills.

Overview

Deep learning model that predicts job categories and calculates match scores between jobs and user skills. Built with CNN-LSTM architecture for mobile deployment.

Features

• Job Category Prediction: Classifies jobs into 20+ categories
• Match Score Calculation: Rates job-skill compatibility (0-1)
• Mobile Optimized: Exports to TensorFlow Lite for Flutter
• Offline Capable: Runs locally in mobile apps

Installation

pip install tensorflow numpy pandas scikit-learn joblib

Usage

1. Train & Export Model

from job_matching_model import JobMatchingSystem

# Train model
job_matcher = JobMatchingSystem()
job_matcher.train()

# Export to TensorFlow Lite
converter = tf.lite.TFLiteConverter.from_keras_model(job_matcher.model)
tflite_model = converter.convert()

# Save .tflite file
with open('job_matching_model.tflite', 'wb') as f:
    f.write(tflite_model)

2. Add to Flutter App

Place the exported model in Flutter assets:

# pubspec.yaml
flutter:
  assets:
    - assets/models/job_matching_model.tflite
    - assets/models/tokenizer.json
    - assets/models/skill_encoder.json

3. Flutter Integration

import 'package:tflite_flutter/tflite_flutter.dart';

class JobMatchingService {
  late Interpreter _interpreter;
  
  Future<void> loadModel() async {
    _interpreter = await Interpreter.fromAsset('assets/models/job_matching_model.tflite');
  }
  
  Map<String, double> predict(String jobDescription, List<String> userSkills) {
    // Implementation for running inference
  }
}

Model Performance

• Accuracy: ~85-90%
• Model Size: ~2-5MB (TFLite)
• Inference Time: <100ms on mobile

Supported Categories

Software Developer, UI/UX Designer, Teacher, Healthcare Worker, Marketing Specialist, Chef, Veterinarian, Social Worker, Personal Trainer, Construction Manager, Writer, Childcare Provider, Event Coordinator, Landscaper, Travel Guide, Actor, Singer, Game Developer, Sports Coach, Volunteer Coordinator, you can add as many as you want!

Files

• job_matching_model.py - Training script
• job_matching_model.ipynb - Jupyter notebook
• job_matching_model.tflite - Exported mobile model
• tokenizer.json - Text preprocessing
• config.json - Model configuration

Contact

Email: asakuraku000@gmail.com