🕵️‍♀️ Fake Job Detector – AI-Powered Job Scam Identifier

Fake Job Detector is an AI-powered tool that helps job seekers identify fraudulent job postings. With a trained machine learning model and integrated Gemini-2.5-Flash-Lite LLM, this tool not only classifies jobs as Real or Fake but also explains why a job is likely fraudulent helping users stay safe from scams demanding personal data or money.

Screenshots

How It Works – Step-by-Step

Dataset Collection

• Used dataset from kaggle.
• Contains job descriptions labeled as fake or real.

Data Cleaning

• Dropped unnecessary columns (e.g., job ID, salary range).
• Replaced null values with blank strings for consistency.

Exploratory Data Analysis (EDA)

• Used RandomUnderSampler (from imblearn) to handle data imbalance – there were too many real jobs vs. fake.

• Visualized:

  • Number of job postings per country (USA dominated).
  • Job experience distribution (Entry-level had the highest share).
  • Most common words in real vs. fake job descriptions using word clouds.

most common words used in real job descriptions

most common words used in fake job descriptions

Text Preprocessing

• Removed stopwords using NLTK to clean up textual noise.
• Normalized and tokenized the text data.

Train-Test Split & Vectorization

• Split data into train and test sets.
• Used CountVectorizer to convert job descriptions into numerical form (Document-Term Matrix).

Model Building & Evaluation

• Trained using Naive Bayes – known for strong performance on text data.

• Also tried Decision Tree, but Naive Bayes + CountVectorizer gave better accuracy.

• Evaluation Metrics:

  • Accuracy
  • Precision
  • F1 Score
  • Support

Model Saving

• Saved the trained model and vectorizer using pickle for web deployment.

LLM Integration (Gemini-2.5-Flash-Lite)

• Integrated Google’s Gemini Flash model to explain why a job is real or fake.
• After prediction, users can click the "Explain Why" button to get an AI-generated reasoning.

Tech Stack

Component Tool/Library

Frontend HTML, CSS, JavaScript
Backend Flask
ML & NLP Scikit-learn, NLTK, CountVectorizer, Imbalanced-learn LLM Gemini-2.5-Flash-Lite via Google Generative AI API
Model Naive Bayes Classifier

Use Cases

• Freshers & students verifying internship offers.
• Job seekers avoiding scam job listings.
• Platforms enhancing job post moderation.
• Career counselors verifying job leads before sharing.

Run Locally

1. Clone the repo:

   git clone https://github.com/yourusername/fake-job-detector.git
   cd fake-job-detector

2. Install requirements:

   pip install -r requirements.txt

3. Run the Flask server:

   python app.py