Finding a job online has become much easier thanks to modern job platforms, but there’s a downside—fake job postings are everywhere. These fraudulent listings aren’t just annoying; they can lead to real problems like stolen personal information and financial losses for unsuspecting job seekers. Our study tackles this issue head-on by developing a smart system that uses deep learning to spot fake job postings. We worked with a dataset called the Fake Job Posting Prediction dataset, which includes both legitimate and fraudulent job listings, to train our detection model. While previous research has shown promising results using models like BERT and RoBERTa for text analysis, we wanted to push things further. We created a hybrid approach that combines three powerful technologies: GPT-2, XLNet, and LSTM networks. Here’s how they work together: GPT-2 and XLNet help the system understand the context and language patterns in job postings, while LSTM captures the sequential patterns that often give fraudsters away. This combination helps us catch fake postings more accurately while keeping false alarms to a minimum. We tested our model using standard metrics—accuracy, precision, recall, and F1-score—and the results were encouraging. Our hybrid model outperformed existing methods, offering better protection for people searching for jobs online. At the end of the day, this research is about making the job search safer. We want job seekers to feel confident that the opportunities they’re applying for are legitimate, helping to build more trust in online recruitment platforms.