Detecting Fake Job/Internship Postings

Inspiration

It started when a close friend of mine was targeted by a fake internship posting. He spent weeks completing “tasks,” sharing documents, and communicating with what turned out to be a completely fabricated company profile. The experience wasn’t just embarrassing — it cost him time, energy, and trust. Watching that happen made me realize how dangerously easy it is to fall for fake job or internship listings, especially for students who are just starting out.

That was the moment I decided I wanted to build something that protects people like him — and all of us — from scams that hide in plain sight.

What We Learned

While working on this project, we learned: How widespread fake recruitment scams actually are. What patterns consistently appear in fraudulent postings (odd emails, unrealistic promises, mismatched domains, generic job descriptions). How users think about trust online — most people rely on vibes, not verification. How important early detection is: a single warning at the right time can prevent huge losses. We also learned how to translate these behavioral insights into technical checks, like domain validation, NLP-based scam detection, and cross-platform verification.

How We Built It

We approached the build in clear layers: Data collection — We gathered examples of both verified and fake job postings to identify patterns.

Feature extraction — Using simple NLP and heuristics, we mapped signals like: [+] suspicious domain names [+] generic job descriptions [+] unrealistic pay or promises [+] mismatched emails [+] reused scam text across different sites

Detection engine — We built a lightweight scoring model combining: [+] rule-based checks [+] text classification [+] crowdsourced flags

Browser extension — To minimize friction, we designed a one-click overlay that shows: [+] a verified/flagged badge [+] risk factors (score based) [+] reasons for suspicion (links)

User testing — We tested with students who frequently apply for internships and refined the UI to be clear, non-technical, and fast.

Built With

• browser-extension-apis
• canva
• figma
• javascript
• mongodb
• node.js
• python
• react
• scikit-learn
• spacy
• typescript
• user-reports
• whois