About the challenge

Hiring has a trust crisis on both sides of the funnel, and AI cracked it open. Recruiters can't tell which applicants are real; candidates can't tell which employers are. This is your one day to build something that restores it.

Trust in the Hiring Funnel is a full-day, in-person hackathon co-hosted by localhost:nyc, Integral Recruiting (IRD), and NYU. What makes it different: you're not building in a vacuum. Real talent-acquisition practitioners from IRD's ~350-member SAI community are in the room all day to ground your build in how hiring actually works — and to tell you whether a recruiter would trust what you made.

We're looking for working tools that solve a narrow slice of the problem and fit real hiring workflows — a screening/triage layer that plugs into an existing applicant tracking system, or a verification tool that makes employer impersonation harder to pull off. We are not looking for broad concept decks, "AI for HR" platforms that do everything, or prototypes that don't run. Demonstrable beats ambitious.

Get started

Talent acquisition runs on a basic assumption: that the person applying is who they say they are, and the company reaching out is who it says it is. Generative AI broke both halves at once. Pick a side of the funnel — or the seam between them — and rebuild trust.

Starting points to spark ideas (inspiration, not requirements):

 

  • Separate real humans from AI noise — a screening layer that flags manipulated or machine-generated applications without false-flagging honest candidates who just used AI to polish a resume.
  • Detect sameness — when 500 applications converge on the same AI-generated phrasing and structure, surface the genuine signal underneath.
  • Catch prompt injection — spot hidden text designed to manipulate resume-screening LLMs before it poisons the funnel.
  • Verify the employer — a lightweight "trust signal" for a careers page or offer letter that lets a candidate confirm an offer is legitimately from the company.
  • Flag impersonation early — detect look-alike domains spun up to send fake offers or run deepfake recruiter calls.
  • Surface the right candidates fast — a triage tool that fits inside an existing recruiting workflow rather than replacing it.

Requirements

What to Build

A working tool that raises trust somewhere in the hiring funnel — built during the event. It should solve a narrow, specific slice of the problem and fit the way hiring actually happens: a triage/screening layer that plugs into an existing applicant tracking system, or a candidate/employer verification tool. Prototypes must be demoed live at the end of the day — no video-only submissions. A real, running solution to a small problem beats a broad concept that doesn't execute.

What to Submit

On your Devpost project page, include:

 

  • Project name and a description
  • What you built and why — the specific slice of the trust problem you attacked
  • Tools and technologies used
  • Team member names and roles
  • Link to your live demo or working prototype (repo or hosted link)

Hackathon Sponsors

Prizes

1 non-cash prize
Winner
1 winner

Credit Prizes - TBD

Devpost Achievements

Submitting to this hackathon could earn you:

Judges

Somya Gupta

Judging Criteria

  • Design (out of 5)
    Is it usable and coherent? Would a recruiter or candidate actually understand and trust the interface in the moment they need it?
  • Uniqueness of Idea (out of 5)
    Does it attack the trust problem from an angle that isn't already a crowded field of look-alike tools? Original framing of the signal or the verification approach counts here.
  • Technical Execution (out of 5)
    Does it run? We reward a real, demonstrable solution to a narrow slice over a broad concept that doesn't work. Bonus for fitting into a real hiring workflow (e.g., plugging into an ATS) rather than replacing it.
  • Real-World Impact (out of 5)
    Does it solve a big real world problem?

Questions? Email the hackathon manager

Invite others to compete

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.