PROOF//HIRE

What Inspired Us

A friend almost handed her Social Security number to a scammer.

The email looked real — Amazon logo, job ID, a link to apply. The only problem? The domain was careers-amazon-hr.net. Not amazon.com. One substring she'd never have caught.

When we looked into it, the numbers were worse than we expected. $660 million stolen from job seekers in 2023. Employment scams ranked the #1 consumer scam in America. And AI is making it cheaper every day to register a look-alike domain, clone a voice, or generate a fake offer letter.

So we asked a simple question: when a candidate gets a suspicious email, where do they go to check?

The answer was nowhere. The entire trust infrastructure — background checks, brand protection, ATS screening — is built for employers. Nobody builds for the person actually getting scammed.

We decided to build that tool in a day.

What It Does PROOF//HIRE is a 10-second trust check for job seekers.

Paste a suspicious URL, email domain, or company name. Get an instant answer:

✅ Looks Legitimate — domain matches, company verified ⚠️ Verify Carefully — something doesn't add up 🚨 Likely Impersonation — this is a look-alike, here's the real one Plus: the official careers page link and clear next-step guidance.

We don't score candidates. We don't make hiring decisions. We verify employers.

How We Built It Four hours. Next.js on Vercel. One API route. One external service. No database.

The core is a verification engine that combines two layers:

Structural analysis — catching the patterns human eyes miss in look-alike domains Live web search — confirming the company exists and checking for known scam reports Both layers run server-side in under 10 seconds. The output is a single trust signal with evidence and an actionable link.

We deliberately cut scope. No WHOIS. No certificate transparency. No email header analysis. No deepfake detection. Every feature was evaluated against one filter: does this help the candidate right now? If not, it went on the roadmap.

Challenges The "use all the tools" trap. We had access to three hackathon tools. We used one. The others added complexity without improving the 10-second verification flow. Constraint was the hardest and best decision we made.

False positives. Flagging a real company as suspicious would destroy trust in the tool. We built conservative thresholds and always surface the official URL — even if we're uncertain, the candidate ends up in the right place.

Five hours. Not enough for a production tool. More than enough for a working proof of concept. We rehearsed the demo four times and deployed twice.

What's Next The prototype verifies domains. The roadmap verifies people:

Face verification — is the recruiter on this video call the person in their LinkedIn photo? Voice detection — is this voice real or AI-generated? Employer badges — companies pre-verify so candidates see a trust signal before they even check. The same question — does this entity match what it claims to be? — applies to domains, faces, voices, and companies. We start with domains. We don't end there.

Founder: Ramesh Lekshmynarayanan Hackathon Team Members: Mohammed Aroui, Ahmad Nawaz

Powered by Green Catapult Labs — NYU Hackathon, September 19 2026.