About the Role

As AI agents move from answering questions to taking action across the internet, the standards that govern identity, delegated authority, discovery, communications, agentic-commerce and payment interfaces, agent-facing web access, agent-specific provenance, and auditability will determine whether agents can operate safely, interoperably, and with clear accountability. OpenAI is looking for an Agent Standards Manager to lead execution of our external technical standards strategy for a defined set of agent standards workstreams.

This role will work closely with Agent Security, Applied and Platform teams, Product, Legal, GRC, Global Affairs, and GTM to turn OpenAI's technical architecture and controls into credible specifications, protocol profiles, assurance criteria, and implementation guidance-and to bring emerging external requirements back into product and security roadmaps before they become blockers. The positions developed here are technical standards positions, not broad public-policy positions.

This is a builder's role as much as an external-facing role. You will execute a prioritized standards plan, author and negotiate protocols, build coalitions, secure the right internal approvals, and represent OpenAI in assigned technical forums. You will also help create repeatable processes that let the company shape agent standards without depending on a single person or relying on technical teams to carry external standards work on top of their core jobs.

The Agent Standards Manager is an integrator, not the owner of every underlying technical system. Security, Product, Engineering, and other designated technical teams retain ownership of architecture, threat models, controls, product decisions, approvals, and implementation. This role owns workstream planning, standards translation, drafting, negotiation, external representation within approved mandates, commitment traceability, and implementation handoff.

In this role, you will:

• Own a focused workstream plan for core agent identity, discovery, and delegated-authorization standards, with adjacent agentic-commerce, web-access, provenance, and auditability work assigned through portfolio prioritization.
• Recommend where OpenAI should lead, shape, monitor, oppose, or decline to participate within the assigned workstream, based on technical maturity, governance quality, strategic leverage, implementation readiness, and company priorities.
• Develop technical standards positions and normative contributions that preserve distinctions among the human or service principal, agent, workload or runtime, task, session, resource, and policy decision.
• Shape requirements for scoped delegation, attenuation, step-up approval, revocation, credential custody, runtime binding, discovery, interoperability, agent-specific provenance, and auditable action trails.
• Embed with Agent Security and Applied or Platform teams to translate internal architecture and controls into technically credible specifications, profiles, white papers, and assurance evidence without overstating maturity or deployment status.
• Lead review and approval paths across Product, Security, Engineering, Legal, GRC, Global Affairs, and GTM; maintain clear mandates, red lines, owners, dependencies, and external commitments.
• Represent OpenAI in priority venues and coalitions, such as IETF and OAuth-related work, FIDO, OIDF, W3C, AAIF, and relevant NIST or ISO efforts; build trust with peer companies, standards operators, and public-sector stakeholders.
• Bring external standards intelligence back into product, security, compliance, and go-to-market decisions early enough to shape roadmaps and avoid retrofit, launch, procurement, or market-access friction.
• Contribute reusable playbooks, decision records, contribution patterns, and handoffs that reduce key-person risk across the assigned workstream.

You might thrive in this role if you:

• Have strong experience in technical standards, internet protocols, identity and access, security architecture, applied cryptography, or closely related product or infrastructure work.
• Have authored, edited, or materially shaped normative specifications, profiles, drafts, or conformance requirements in serious technical standards or open-source governance settings.
• Can reason clearly about OAuth and OIDC, FIDO or WebAuth, workload identity, HTTP signatures, delegated authority, credential lifecycle, revocation, provenance, and auditability; you do not need to be expert in every layer.
• Understand that agent trust is a composed stack rather than a single credential, and can separate principal, agent, runtime, task, session, resource, and policy claims.
• Can distinguish adopted standards, stable extensions, active drafts, working-group proposals, and exploratory ideas-and know how that maturity should change OpenAI's posture.
• Earn trust with engineers and security leaders while also communicating crisply with lawyers, product leaders, executives, policymakers, and external standards participants.
• Have strong judgment in ambiguous, high-stakes environments and know when to lead, compromise, oppose, escalate, or walk away.
• Write clearly, negotiate constructively, and are comfortable representing OpenAI externally with discretion.