VeriHire

About the Project

Today's hiring and onboarding systems force candidates to hand over highly sensitive documents — passports, visas, work permits, immigration records, transcripts, and personal identifiers — to multiple recruiters, HR systems, vendors, and third-party platforms. While companies need to verify eligibility and compliance, candidates often have no visibility into where their data is stored, who can access it, or how long it is retained.

We were inspired by a simple question:

Why should proving a fact require exposing the entire document behind it?

Current hiring systems operate on an “upload everything” model. That creates unnecessary privacy risks, especially for international students, immigrants, and visa-sponsored employees whose immigration status is deeply sensitive information. We discovered real-world cases of HR platforms leaking employee data, immigration records being mishandled, and employers improperly collecting or storing work authorization documents. At the same time, existing onboarding tools mostly focus on compliance workflows — not candidate privacy.

Our project rethinks employment verification using privacy-preserving cryptography and AI.

Instead of uploading raw identity documents, candidates generate cryptographic proofs that verify only the information an employer actually needs:

• Is this person authorized to work?
• Does the visa expire before the employment term?
• Is sponsorship required?
• Is the credential authentic?

The employer receives the verification result — not the underlying passport, visa, or immigration document.

How We Built It

We built a privacy-first verification platform combining:

• Midnight confidential smart contracts for on-chain proof generation and selective disclosure verification — the core of what makes this trustless and privacy-native
• AI-powered document analysis to validate authenticity and extract required claims
• Encrypted credential storage so raw documents never touch the verification layer
• Role-based access controls and audit logging for compliance visibility

Our architecture separates:

1. Document ingestion and proof generation
2. Midnight smart contract verification logic
3. Employer-facing verification APIs
4. Privacy-preserving credential sharing

The frontend focuses on a clean onboarding experience for candidates and recruiters, while the backend handles verification workflows and cryptographic proof validation.

For this hackathon MVP, we demonstrate the full flow using résumé credentials — degree verification, GPA thresholds, and skill claims — as a proof of concept for the broader work authorization use case. The same Midnight proof infrastructure applies directly to passport and visa verification.

Challenges We Faced

The biggest challenge was balancing:

• privacy
• compliance
• usability

Employment onboarding is heavily regulated, and many existing systems assume full document collection is necessary. Designing a workflow that minimizes exposure while still satisfying verification requirements required extensive research into immigration processes, I-9 compliance, HR systems, and identity verification standards.

Another challenge was translating advanced cryptographic ideas — specifically Midnight's confidential smart contract model — into a product experience that recruiters and candidates could realistically understand and use during a hiring process.

We also had to carefully define our scope. Existing HR vendors already solve document management and compliance workflows, so we focused specifically on the privacy gap: reducing unnecessary exposure of candidate identity and immigration data.

What We Learned

This project taught us that privacy is often treated as an afterthought in hiring systems. Most platforms optimize for convenience and compliance, but not for minimizing data exposure.

We also learned that:

• Privacy-preserving identity verification is technically possible today
• HR and immigration workflows still rely heavily on centralized document storage
• Candidates have very little control over their sensitive onboarding data
• AI and cryptography together can enable entirely new trust models

Most importantly, we learned that building responsible AI systems is not just about model capability — it’s about designing systems that reduce unnecessary access to sensitive human data in the first place.

Built With

• compact
• docker
• fastapi
• gemini
• jwt
• midnight
• next.js
• postgresql
• python
• react
• sql
• tailwind
• typescript