AI Deployment Engineer, Agent Enablement

OpenAI•San Francisco, CA

18h•$197,000 - $280,000•Hybrid

About The Position

The Agent Enablement AI Deployment Engineering (ADE) team works across engineering, product, design, partnerships, and strategic customers to grow an open ecosystem of agent-enabled sites and services. We help partners adopt the OpenAI tech stack related to identity, permissioning, agent-auth primitives so users can safely connect ChatGPT and Codex to the tools, services, and workflows they already use. Our team also works with external partners on defining the standards for agent access, marketplace offerings as well as other agent enablement initiatives to ensure users of ChatGPT and Codex go from intent to task completion seamlessly. We are looking for an AI Deployment Engineer to help strategic partners design, build, validate, launch, and operate agent enablement integrations across web applications, connectors, APIs, CLIs, MCP servers, and developer tools. This is a hands-on, partner-facing product engineering role for someone who can contribute to the platform itself, lead sophisticated technical engagements, and turn ambiguous identity and agent-workflow requirements into secure, production-ready integrations. You will work across partner product and engineering teams and OpenAI’s product, engineering, design, partnerships, legal, policy, security, support, and go-to-market teams. You will identify high-value user journeys, choose the right integration path, prototype and review architectures, write code, run evaluations and dogfood, trace failures end to end, guide launch and rollout, and support post-launch iteration. The best person for this role moves fluidly between full-stack code, OAuth/OIDC and identity systems, product judgment, project leadership, and clear communication with engineers and executives. This role is a fit for a product-minded engineer who wants to stay close to users and partners while going deep on authentication, permissions, reliability, safety, and developer experience. The principle objective is to help partners ship low-friction, trustworthy agent experiences that work across real-world account states, enterprise constraints, and user expectations. This role is based in our San Francisco office. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees.

Requirements

4–6 years of professional software engineering or AI engineering experience and are strong enough technically to contribute to the platform itself while still enjoying hands-on coding.
Operated in a customer or partner facing roles in scoping projects, building MVPs, presenting trade-offs in solution design and architecture, understanding user value drivers as well as scoping and executing on tight implementation plans.
Built and operated production full-stack products, APIs, backend services, developer platforms, connectors, or integrations, and can reason across frontend, backend, auth, data models, reliability, privacy, and UX constraints.
A strong understanding of how plugins and connectors work with coding agents, integrating MCP servers, using CLIs and APIs to access external tools.
Can work comfortably with external engineers, product leaders, and executives, and translate partner feedback into crisp technical plans and product improvements.
Are comfortable with ambiguity and rapidly changing conditions, and can turn high-level ideas into runnable prototypes, API contracts, technical specs.
Balance urgency with judgment and can keep complex partner launches moving while staying precise about the details that matter.

Nice To Haves

High-level understanding of the identity stack, including how OAuth works.
Experience designing or operating APIs, webhooks, schemas, SDKs, CLIs, MCP servers, tool-calling systems, or other developer-facing integrations.
Prior customer-facing or partner-facing engineering experience, including leading technical calls, navigating ambiguity, and helping external teams ship production software.
Strong familiarity with AI products, LLM APIs, ChatGPT or Codex surfaces, coding harnesses, developer platforms, agent workflows, or ecosystem products.
Experience with evaluations, staged rollouts, observability, incident response, launch operations, and turning production learnings into durable platform improvements.

Responsibilities

Own the technical partner journey for priority agent enablement integrations—from use-case selection and readiness assessment through architecture, prototype, implementation, evaluation, launch, rollout, and ongoing maintenance.
Help partners choose and implement the right integration path across browser sign-in, connector-initiated OAuth, and agentic account linking or provisioning, with clear user journeys and safe fallback behavior.
Write production and sample code, build reference implementations and test harnesses, and create the technical guidance, integration checklists, evaluations, and debugging tools that move partners from concept to production.
Debug identity and agent workflows end to end: user interface state, browser redirects, PKCE/OIDC transactions, token exchange and validation, account mapping, connector callbacks, CLI or MCP handoffs, latency, retries, rate limits, logs, traces, and metrics.
Review partner architectures and implementation plans for API contracts, scopes and permissions, consent, terms acceptance, and account policy, secret handling, data boundaries, privacy, reliability, and long-term maintainability.
Run hands-on evaluations, dogfood, launch-readiness reviews, staged rollouts, and post-launch investigations; turn findings into concrete fixes rather than one-off workarounds.
Contribute targeted improvements to ChatGPT, Codex, and the Agent Enablement platform, including identity protocols, APIs, SDKs, docs, examples, internal tooling, partner-debugging workflows, launch guardrails, and user-facing consent or control experiences.
Work closely with product, engineering, design, partnerships, legal, policy, security, support, and go-to-market teams to make partner launches smooth, safe, and repeatable.
Bring structured signals from partners back to product and engineering, and turn recurring integration patterns into platform requirements, reference architectures, playbooks, and developer guidance.