Census Networks treats hiring as a supply-chain control. Our job is to make every employer able to reason about candidate risk the way modern security teams reason about every other control.
The capabilities to detect employment fraud already exist as point tools, owned by different teams, none of them reasoning about the joint pattern. State actors and financial fraudsters are very good at finding that seam. We are building the orchestration layer that closes it, and we are doing it in the open, with explainability, human-in-the-loop, and shared-defense by design.
A sophisticated adversary does not attempt to fool any single tool. The known facilitator playbooks include coaching on which background-check vendors specific employers use, which OSINT signals to plant in advance, and which interview platforms can be defeated by which face-swap stack. The adversary is reasoning about the joint defense and finding the seam. The defense has to do the same. That is the company in one sentence.
Census draws on advisors from across security, HR, adversarial ML, and policy. The full advisory roster will be announced as the development-partner phase progresses.
Judgments are formed from joint patterns, not single artifacts.
No single model dominates. An adversary cannot defeat the platform by defeating one component.
Candidates are monitored from application through continuous post-hire; the moment of hire is not a hand-off out of risk visibility.
When signals disagree, the platform errs toward caution and routes to human review.
Every score decomposes into the evidence that produced it. No black boxes in the decision path.
We are running pilots with a small set of security vendors and enterprise employers, refining the platform against live adversarial conditions, and convening the working group around our cross-industry signal-sharing proposal. The roadmap is real, the calendar is real, and we publish what we measure.