ScenarioInternational frontier AI verification and emergency coordination

Frontier Compute Verification Specialist

Also called AI Compute Treaty Verifier

A Frontier Compute Verification Specialist would design checks that show whether compute-related governance commitments are being kept. The role depends on how AI develops and on whether states authorise international assurance. It is not a role companies are hiring for now.

What they do

If international assurance institutions emerged, this person would turn a governance commitment about compute into a check that could actually be run. They would work out what evidence could show compliance, who could collect it and under what legal authority. They would write verification protocols and test them by asking how a determined party might evade them. They would sit with lawyers, inspectors, hardware specialists and diplomats, and explain technical limits in plain language. They would also say clearly when a commitment cannot be verified.

What a strong hire delivers

In the first months, a strong hire would deliver a draft verification protocol for one defined commitment, with its assumptions stated. They would add an evasion and feasibility analysis that ranks the weakest points. They would also leave a clear list of what legal authority and tooling the protocol needs, and a shared view among colleagues of what counts as verified.

How you can tell they are good

Ask them to take a commitment about compute and show how they would verify it. A strong candidate states the constraints first, separates what is observable from what is claimed, and attacks their own design. They are open about residual uncertainty. Be wary of anyone who promises certainty, ignores legal limits, or treats verification as a purely technical exercise.

Why it is hard to fill

If this happened, the people would be scarce because the work sits between disciplines that rarely meet. Technical depth in compute is common in one world, and inspection or treaty experience in another. Few people hold both. Titles would vary or not exist, and there is no established career route, so a recruiter would be judging potential from adjacent work.

Where to find them

Today they would come from international governance and arms control verification, inspection and audit work, assurance and compliance in safety-critical industries, and emergency coordination. Others would come from AI governance and risk teams, policy groups close to frontier labs, and technical staff in hardware or infrastructure who have worked with regulators or standards bodies.

Interview questions for a Frontier Compute Verification Specialist

  1. Describe a time you designed or ran a check that others had to trust. What made the evidence credible?

    A strong answer names the claim being tested, the evidence chosen and why others accepted it. It admits limits and disputes. Worry if they describe only tools, never the audience who had to rely on the result.

  2. Take a commitment to cap compute use. What would you need to know before proposing any verification method?

    They should ask about definitions, scope, legal authority, access and the parties involved. They separate measurable from claimed quantities. Worry if they jump straight to a favourite technical method.

4 more questions, with what a strong answer covers, are in the Role Kit.

Job description you can post

We are building capability to check whether commitments about advanced AI compute are being kept. You would design verification protocols that are technically credible and workable within legal limits, and you would work with specialists across law, inspection and policy.

What you will do

  • Design verification protocols for defined compute-related commitments
  • Analyse how protocols could be evaded and how feasible they are
  • State assumptions, limits and residual uncertainty clearly
  • Work with legal, technical and policy colleagues on what is enforceable
  • Advise on the tooling and authority a protocol requires
  • Write plain summaries for non-technical decision makers

What you will bring

  • Demonstrated work designing checks, audits or inspections others relied on
  • Working knowledge of how compute is used in frontier AI development
  • Experience analysing how a control could be defeated
  • Clear writing for mixed technical and non-technical readers
  • Experience working across legal, policy and technical teams
  • Sound judgement about what can and cannot be verified

Nice to have

  • Experience in international or treaty-based settings
  • Background in emergency coordination or incident response
  • Familiarity with hardware or data centre operations

Free to use and adapt. If it saves you time, a link back to this page is appreciated.

Depends on a scenario

International frontier assurance institutions

What AI would need to do
Frontier-system risks create incentives for cross-border verification and coordination.
What organisations would need to decide
States or other legitimate actors authorise workable international assurance mechanisms.
What would rule it out
No agreement, no verification feasibility or no enforcement capacity.

We give no date. The role becomes real if both conditions hold.

Skills that matter

  • Institutional and international coordinationAdvancedAlign actors and processes across jurisdictions
  • AI governance and assuranceAdvancedAssign accountability and verify lifecycle controls
  • AI risk managementWorkingIdentify, assess, treat and monitor AI risks
  • Frontier scenario analysisWorkingExplore conditional outcomes without claiming predictions

Other roles in international frontier AI verification and emergency coordination

ScenarioInternational Frontier AI Verification InspectorAn International Frontier AI Verification Inspector would assess compliance with an international agreement on advanced AI. The role depends on how…ScenarioGlobal AI Incident Coordination OfficerA Global AI Incident Coordination Officer would coordinate cross-border response to serious frontier AI incidents. Whether the role exists depends…ScenarioInternational AI Assurance Accreditation DesignerAn International AI Assurance Accreditation Designer would set how assurance providers are judged competent and independent within a possible…

Related roles in other families

ScenarioAutonomous Agent Identity Governance ArchitectAutonomous agent institutions and economic coordinationAn Autonomous Agent Identity Governance Architect would design how agents prove who they are and whose authority they carry. It depends on how AI…ScenarioAgent Economy Assurance AnalystAutonomous agent institutions and economic coordinationAn Agent Economy Assurance Analyst would assess systemic risk where autonomous agents transact and coordinate. The role depends on how AI develops…ScenarioMachine-to-Machine Dispute Process DesignerAutonomous agent institutions and economic coordinationA Machine-to-Machine Dispute Process Designer would build the routes for resolving disputes when autonomous agents act for people or organisations…ScenarioAgent Collective Action Governance SpecialistAutonomous agent institutions and economic coordinationAn Agent Collective Action Governance Specialist would design rules for groups of AI agents acting together at scale. It depends on how AI develops…

How we wrote this

Role definitions are our own editorial work, checked against occupational classifications, standards, research and employer pages. Titles vary by company, and this is a description of the work, not a job advert. Scenario roles describe responsibilities that depend on stated conditions; they are not forecasts.

← International frontier AI verification and emergency coordinationAll new job families

Start a search

Hiring a Frontier Compute Verification Specialist?

Tell us the work. We find the people already doing it, whatever their job title says, and you meet them already interviewed for your role.