AI Student Builds Chrome Extension to Combat 90% Ghost Internship Rate

90% of the internships I applied to weren’t real. So I’m building a way to expose them.

AI student Quratulain Nayeem is developing a Chrome extension to identify “ghost jobs” on LinkedIn. The project addresses a market where an estimated 90% of listings in certain sectors are used solely for data collection or talent pipelining rather than active hiring.

Why This Matters

The technical reality of the job market involves a proliferation of “ghost listings” where companies post roles they do not intend to fill to gauge salary expectations or build talent pipelines. This creates a massive inefficiency for developers who optimize resumes for non-existent roles, highlighting a significant gap between public job board data and actual hiring signals.

Key Insights

• 90% of internship listings in certain sectors are estimated to be ghost listings used for data harvesting and talent pipelining.
• Legitimacy scoring aggregates five signal categories: Company Legitimacy, Recruiter Credibility, Posting Behavior, and On-Device ML.
• ONNX Runtime Web is utilized for local text classification to ensure privacy-first, low-latency performance within the browser.
• Manifest V3 extensions provide a ‘Truth Layer’ directly on LinkedIn, bypassing the restrictions of enterprise-locked APIs.
• Vercel Edge functions are implemented to handle the backend logic for real-time credibility calculations and domain cross-referencing.

Practical Applications

• LinkedIn Job Seekers: Using an extension-based ‘Truth Layer’ to filter out illegitimate listings requiring payment (e.g., ₹1,499 AI internships).
• Recruiters: Maintaining credibility by ensuring active hiring signals align with posting behavior to avoid being flagged by legitimacy scoring systems.

References:

• https://dev.to/qurotz/90-of-the-internships-i-applied-to-werent-real-so-im-building-a-way-to-expose-them-2jnm