job_scam_detector

• 

  postman sending job ai detecting its fake

• 

  server running

Inspiration

In the gig economy, freelancers and job seekers often fall victim to fake job postings, leading to wasted time, financial loss, and security risks. We wanted to create a smart AI tool that helps people detect scam jobs instantly and avoid fraud.

What it does

Scans Job Listings

Takes job titles and descriptions as input. Analyzes job details using an AI model trained on real and fake job data. Predicts Fake Jobs

Classifies jobs as "Real" or "Fake" based on patterns in the text. Helps users avoid scams and fraudulent job postings. Works with a Browser Extension

Can be integrated into Google Chrome to analyze jobs directly on job boards. Provides instant feedback on whether a job is safe to apply for.

How we built it

Data Collection & Processing – Gathered real and fake job listings, cleaned the data, and converted text into numerical form using TF-IDF vectorization.

AI Model Development – Trained a Deep Learning model (TensorFlow/Keras) to classify jobs as Real or Fake, then saved it for deployment.

API Development – Built a Flask API to handle job description inputs, preprocess them, and return predictions.

Browser Extension – Created a Chrome Extension that interacts with the API to flag fake jobs in real time.

Deployment & Testing – Hosted the API on Render, fine-tuned the model, and improved accuracy based on real-world job postings.

✅Tech Stack: Python, TensorFlow, Flask, Pandas, Scikit-learn, Chrome Extension APIs, Render.

Final Outcome: A real-time AI-powered tool to help job seekers detect scam job listings!

Challenges we ran into

🔹 Web Scraping Limitations: Google Jobs’ structure changed, requiring frequent adjustments. 🔹 Data Imbalance: The dataset contained more real jobs than fake ones, affecting model accuracy. 🔹 Model Optimization: Initially, our model was overfitting; we improved it by tuning hyperparameters. 🔹 API Errors: Encountered 404 errors and shape mismatches during API testing, which we debugged. 🔹 Deployment Issues: Faced compatibility problems with TensorFlow on cloud servers, requiring a lightweight alternative.

Built With

• css)-storage:-csv-files
• github-pages-(for-documentation)-browser-extension:-chrome-extension-apis-(javascript
• html
• javascript-machine-learning:-tensorflow
• json-deployment:-render-(for-api)
• keras
• numpy-web-scraping:-beautifulsoup
• pickle
• programming-languages:-python
• scikit-learn-data-processing:-pandas
• selenium-api-development:-flask