AI-Based Fake Job Posting Detection Using Multimodel Analysis
Designed and developed a multimodel AI system to detect fraudulent job postings using textual content analysis with BERT, TF-IDF, and sentiment analysis; numerical attributes such as salary plausibility and experience patterns; and URL/domain reputation. Built an ensemble deep learning model using CNN/LSTM with attention, achieving over 92% accuracy with real-time processing under 500ms. Incorporated explainability through SHAP values and deployed the system as a REST API with a web-based dashboard.