OpenAI FDE and adjacent roles: verified jobs and interview guidance

Start with the vacancy, then decide what to prepare

“Customer-facing AI work” spans several different OpenAI vacancies. A Tokyo FDE builds and deploys with customers; a Deployment Lead coordinates the path to business results; an Architect owns an account’s technical strategy. Applied AI Engineers also implement production systems. Choosing by title alone can lead to preparing the wrong evidence.

This is a verified-facts reference, checked on 2026-10-05, with 80 identified facts from 19 official sources. The question articles that follow will be editorial practice, not disclosed interview questions. No candidate anecdotes, private questions or copied vendor images are used. A live job page with an application link is an observation at that date, not a promise of remaining headcount.

Read a fact ID as a stable pointer: the table provides a short claim and a paraphrased evidence area; its adjacent official link gives the source. Exact headings remain in the fact ledger. Source publication and revision dates are unknown unless explicitly stated. The check date neither dates a job’s creation nor makes an article published.

The 13 observed vacancies

Every row below was linked from an official OpenAI job page to Ashby “Apply now” at verification. Observed status for all 13: listed with an application link. The pages did not establish internal level, remaining hiring slots or a guaranteed interview start. The job IDs identify the vacancies; an unlabelled title is not evidence of a particular seniority.

Vacancy / official page	Location	Department	ATS job ID	Level disclosure
Forward Deployed Engineer - Tokyo	Tokyo, Japan	Forward Deployed Engineering	51b17595-3a70-43be-a333-3a3952303284	Unknown
Deployment Lead (DL), FDE - Tokyo	Tokyo, Japan	Forward Deployed Engineering	df514733-46ad-4777-be67-44667bc9fd45	Unknown
Manager, Forward Deployed Engineer - Tokyo	Tokyo, Japan	Forward Deployed Engineering	bef38cc8-b13c-4ca8-8e1c-7d97139efdea	Unknown
Applied AI Engineer - Tokyo	Tokyo, Japan	Technical Success	4762783e-61e1-4d36-af1d-8dee3c5f70e2	Unknown
Applied AI Architect - Tokyo	Tokyo, Japan	Technical Success	57cfe1d3-27b6-4761-909c-fec7829e9227	Unknown
Applied AI Engineer, Codex — Tokyo	Tokyo, Japan	Technical Success	6668417f-6878-4244-940f-9a99ae4ebcb0	Unknown
AI Deployment Manager (Builder) - Tokyo	Tokyo, Japan	Technical Success	bb569a6b-a7e9-4c4d-a78e-d35349b2f9d9	Unknown
Forward Deployed Engineer (FDE) - SF	San Francisco, US	Forward Deployed Engineering	967f94aa-1706-4dba-ac89-bfbc2c38b688	Unknown
Forward Deployed Software Engineer - SF	San Francisco, US	Forward Deployed Engineering	00207abc-49b7-465c-a219-f7c1140f8047	Unknown
Forward Deployed Engineer - Singapore	Singapore	Forward Deployed Engineering	4c078f82-93b9-427c-8fd8-95de198884b4	Unknown
Forward Deployed Engineer - Singapore (Korean Speaking)	Singapore	Forward Deployed Engineering	15ff9f15-c986-4ac3-ac2e-cd0f029afbd5	Unknown
Forward Deployed Engineer - Seoul	Seoul, South Korea	Forward Deployed Engineering	b43339a1-e830-4761-a193-3438f34f9ebb	Unknown
Forward Deployed Engineer - Sydney	Sydney, Australia	Forward Deployed Engineering	9e39d642-6c78-4c21-bc86-49b1773c933c	Unknown

The current Tokyo DL uses the same ATS ID as an older “Technical Deployment Lead - Tokyo” search result. Count it once; the rename date is unknown. The two SF titles have different ATS IDs. Their experience or salary differences do not establish an internal grade ordering.

Company-wide interview guidance, with its scope intact

The official interview guide provides a general example. It does not publish a confirmed Tokyo FDE loop. Its engineering quality criteria are useful preparation context, but do not prove an FDE-specific coding task or interview rubric. AI-use permissions are interview-specific.

Interview guide

Official source — Global. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-I01	Contribution, learning speed and results matter beyond credentials.	What hiring values
OAI-I02	Collaboration, clear communication, feedback and mission alignment matter.	What hiring values
OAI-I03	The published process is illustrative; individual processes vary.	Limits of the general example
OAI-I04	Application review usually takes about a week; prepare background, motivation, goals and team context.	Application and first conversation
OAI-I05	Team-specific assessments may include pair coding, take-homes or technical tests, sometimes several.	Possible assessment formats
OAI-I06	Reasoning and explanation matter; confirm AI permissions for each interview.	Reasoning and permitted assistance
OAI-I07	The general final-stage example is virtual: 4–6 hours, 4–6 interviewers, 1–2 days; SF onsite is discussed.	General final-stage example
OAI-I08	Engineering guidance covers design, code, performance, tests, collaboration and explanation.	Engineering quality dimensions
OAI-I09	A final response usually takes a week; references may be requested. Read the Charter, research and blog.	Decision timing and preparation reading

Tokyo: implementation, delivery and account strategy

The following source groups keep roles separate. Interpretation: use the responsibility you would own to select evidence: personal implementation for FDE/Engineer, planning and measured value for DL, people development for the FDE Manager, account strategy for Architect, developer workflows for Codex, learning/adoption for Builder. This comparison is derived from the job descriptions; it does not assert separate or identical interview processes.

Forward Deployed Engineer - Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-T01	Own discovery through production for strategic customers.	Purpose of the vacancy
OAI-T02	Success includes production use, business impact and evaluation insights informing product/model plans.	Purpose of the vacancy
OAI-T03	Collaborate with Product, Research, Partnerships, GRC, Security and GTM.	Purpose of the vacancy
OAI-T04	Deliver multiple projects from prototypes to stable operation; embed and support adoption.	Work the role delivers
OAI-T05	Balance scope, speed and quality; sequence work and unblock early.	Work the role delivers
OAI-T06	Write code when needed and turn patterns into reusable tools.	Work the role delivers
OAI-T07	Lists 5+ years in engineering/technical deployment, including customer-facing work.	Experience and skills sought
OAI-T08	Write and review production frontend/backend code, including Python or JavaScript.	Experience and skills sought
OAI-T09	Experience building/deploying LLMs or generative models, and understanding their product effects.	Experience and skills sought
OAI-T10	Tokyo hybrid: three office days, relocation support, mainly domestic travel; no percentage stated.	Work location and travel
OAI-T11	Requires fluent spoken/written Japanese and English, an English résumé and interviews in both languages.	Language and application expectations

Deployment Lead (DL), FDE - Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-D01	Translate business outcomes into technical plans across FDEs, researchers and customer engineers.	Purpose of the vacancy
OAI-D02	Define workstream milestones, dependencies and acceptance criteria; protect the critical path.	Work the role delivers
OAI-D03	Cover 0→1, MVP, scale, readiness and change management.	Delivery scope and activities
OAI-D04	Set value hypotheses, baselines and KPIs; measure before/after and report to executives.	Business-value measurement
OAI-D05	Success also includes reliable delivery, rework, reuse, judgment and product influence.	Evidence of success
OAI-D06	Lists 7+ years of customer-facing technical delivery leadership, shipped AI and industry expertise.	Experience and skills sought
OAI-D07	Tokyo three-day hybrid; customer-site engagement typically 25–50%, sometimes higher.	Location and customer-site work

Manager, Forward Deployed Engineer - Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-M01	People manager accountable for team growth, technical results and business value.	Purpose of the vacancy
OAI-M02	Success includes repeatable shipping, clear field feedback and sustainable delivery.	Evidence of success
OAI-M03	Develop people through actionable feedback; organize field staffing and support.	Work the role delivers
OAI-M04	Lists 8+ engineering/delivery years and 2+ managing customer-facing engineers.	Experience and skills sought
OAI-M05	Still expects production frontend/backend coding and review ability.	Experience and skills sought
OAI-M06	Tokyo three-day hybrid, up to 40% mainly local travel; fluent Japanese/English, English résumé, bilingual interviews.	Location, travel and language

Applied AI Engineer - Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-A01	Technical Success role: hands-on use-case selection through production and scale.	Purpose of the vacancy
OAI-A02	Build/debug code, evaluation harnesses, reference implementations and integrations.	Delivery scope and activities
OAI-A03	Make model, agent, retrieval and tool decisions alongside reliability, observability, latency, cost, safety and governance.	Work the role delivers
OAI-A04	Evaluate systematically using representative data, graders, production signals and human judgment.	Experience and skills sought
OAI-A05	Requires personal coding, design, evaluation and debugging contributions beyond coordination.	Experience and skills sought
OAI-A06	Strong Python and fluent Japanese/English; specific industry/OpenAI-product experience is not required.	Experience and skills sought

Applied AI Architect - Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-R01	Own long-term account technical strategy, from pre-sales to adoption and business impact.	Purpose of the vacancy
OAI-R02	Account Director owns commercial strategy; Architect owns technical strategy and the production-value path.	Purpose of the vacancy
OAI-R03	Engaged delivery teams implement projects; Architect retains technical-progress/customer-outcome responsibility.	Work the role delivers
OAI-R04	Prioritize use-case portfolios and designs covering security, privacy, governance and evaluation.	Work the role delivers
OAI-R05	Prototype with APIs/SDKs and judge when to go deep or involve specialists.	Experience and skills sought

Applied AI Engineer, Codex | Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-C01	Codex focus: improve planning, implementation, testing, review and delivery workflows.	Team purpose and workflow
OAI-C02	Evaluate coding systems with representative tasks, automated graders, production signals and developer feedback.	Work the role delivers
OAI-C03	Be an active AI-coding-tool user with a view on productivity improvement.	Experience and skills sought
OAI-C04	Deliver workshops and reusable assets; strong Python and fluent Japanese/English.	Activities and implementation skills

AI Deployment Manager (Builder) - Tokyo

Official source — Tokyo, Japan. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-N01	Post-sales specialist for structured enablement and adoption.	Purpose of the vacancy
OAI-N02	Design/deliver onboarding, advanced training, executive sessions, hackathons and workshops.	Work the role delivers
OAI-N03	Lists 4+ customer-facing/education years and understanding of coding, agents, APIs, RAG, evaluation and fine-tuning.	Experience and skills sought
OAI-N04	Success means confidence, adoption and business value; scale training materials and playbooks.	Delivery scope and activities

An FDE’s end-to-end responsibility does not erase partners’ responsibilities. The DL posting explicitly names FDEs, researchers and customer engineers; the Architect posting distinguishes the Account Director and delivery teams. A candidate should explain their own decisions and the team’s handoffs without claiming sole ownership of every result.

APAC and US: preserve the local conditions

Travel percentages describe different postings, not a worldwide FDE constant. Tokyo FDE gives no percentage; Tokyo Manager gives up to 40%; Tokyo DL discusses customer-site engagement. Singapore, Seoul, Sydney and SF state their own conditions. Travel and three office days are different obligations.

Forward Deployed Engineer (FDE) - SF

Official source — San Francisco, US. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-U01	SF FDE: three-day hybrid, relocation support, up to 50% travel.	Location and travel arrangements
OAI-U02	SF base range: US$185K–300K plus equity; not evidence of Tokyo pay.	Published pay information
OAI-U03	Base varies by location/skills/experience; eligible performance bonus. Internal level is undisclosed.	Pay variability and undisclosed level

Forward Deployed Software Engineer - SF

Official source — San Francisco, US. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-W01	SF FDSWE works with customers/FDEs on API-based custom software and reusable abstractions.	Purpose of the vacancy
OAI-W02	Code alongside customer teams on their infrastructure; iterate on full-stack implementations.	Work the role delivers
OAI-W03	Lists 7+ full-stack years excluding internships, relational databases; founder/early-engineer experience is a plus.	Experience and skills sought
OAI-W04	SF three-day hybrid, up to 50% travel; US$185K–325K base plus equity, individually variable.	Work arrangements and US pay

Forward Deployed Engineer - Singapore

Official source — Singapore. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-G01	Singapore regular FDE: three-day hybrid, relocation support, expected 50% travel.	Location and travel arrangements

Forward Deployed Engineer - Singapore (Korean Speaking)

Official source — Singapore. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-G02	Singapore Korean-speaking FDE: business-level Korean for customers; expected 50% travel.	Customer language and travel

Forward Deployed Engineer - Seoul

Official source — Seoul, South Korea. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-G03	Seoul FDE: three-day hybrid, relocation support, expected 50% travel.	Location and travel arrangements

Forward Deployed Engineer - Sydney

Official source — Sydney, Australia. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-G04	Sydney FDE: three-day hybrid, relocation support, up to 50% travel.	Location and travel arrangements

Dated context and technical references, not hiring instructions

The 2026-05-11 DeployCo announcement supplies dated business context. It does not establish a change to the Tokyo vacancy’s employer. Tomoro was conditional in that announcement; closing has not been established here. This is the only dated organizational event this reference uses; it does not invent a 2024–2026 hiring timeline or turn undated job pages into recent announcements.

OpenAI launches the OpenAI Deployment Company to help businesses build around intelligence

Official source — Global. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-B01	2026-05-11 announcement: DeployCo is a separately operated business, majority-owned/controlled by OpenAI.	Organizational announcement
OAI-B02	Typical engagements start with value diagnosis and prioritized workflows, connecting customer data, tools and controls.	Engagement design in the announcement
OAI-B03	At announcement, Tomoro acquisition was agreed subject to conditions; about 150 FDEs/specialists were expected.	Conditional acquisition statement

OpenAI Frontier | Enterprise platform for AI agents

Official source — Global. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-F01	Enterprise Frontier Program pairs customers with DeployCo FDEs for design, governance and production agents.	Customer program and field support
OAI-F02	Frontier describes agent identity and task-scoped access.	Agent identity and access boundary
OAI-F03	Frontier describes monitoring and detailed logs for visible, auditable agent actions.	Action visibility and logging

Evaluation best practices

Official source — Global. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-E01	Evaluate early/continuously against real task distributions because generative outputs vary.	When and what to evaluate
OAI-E02	Calibrate automated scoring with human judgment; avoid subjective checks or generic benchmarks alone.	Interpreting scores and avoiding weak checks
OAI-E03	Define objectives, dataset, metrics, comparisons and continuous evaluation; test changes and add cases.	Evaluation design sequence
OAI-E04	Evaluate selected tools and arguments, beyond final answers.	Tool choice and argument checks
OAI-E05	Ground multi-agent complexity in evaluation; assess triage and handoffs.	Justifying agents and checking transfers
OAI-E06	Include multilingual, long-context, multi-intent, tool/handoff failures and adversarial cases.	Failure and unusual-input coverage

Enterprise privacy at OpenAI

Official source — Global. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-P01	Covered Business/API data is not used for training by default; explicit opt-in is an exception.	Training-use defaults and exceptions
OAI-P02	Describes AES-256 at rest, TLS 1.2+ in transit and strict access controls.	Encryption and access safeguards
OAI-P03	API inputs/outputs may be retained up to 30 days by default, with exceptions; eligible use cases/endpoints can request ZDR.	API storage conditions and eligibility
OAI-P04	Workspace admins control apps; existing permissions are respected and user authentication is required.	Connected-app control and permissions

Deprecations

Official source — Global. Evidence areas are paraphrased below; exact source headings are retained with fact IDs in the research ledger. Verification: 2026-10-05; publication unknown unless stated. Internal job level remains unknown.

Fact ID	Confirmed statement	Evidence area (paraphrased)
OAI-X01	Evals platform plans read-only on 2026-10-31 and dashboard/API shutdown on 2026-11-30; Promptfoo migration is described.	Dated hosted-platform retirement plan

The 2026-06-03 source date belongs to the Evals platform notice section, not the first publication of the entire Deprecations page. The platform dates in OAI-X01 were still future plans on 2026-10-05. Evaluation as an engineering practice remains distinct from that hosted platform. Privacy statements apply to their named products, endpoints and exceptions; “not used for training” is not a claim of zero storage. These references support technical preparation; none discloses an interview question.

A responsibility map for selecting evidence

This is an original conceptual reading map, derived from the Tokyo job responsibilities. Arrows mean a dependency between work products; they do not depict a mandatory company workflow, reporting line or interview sequence.

1. 1Customer workflow and baseline
2. 2DL plan and acceptance criteria
3. 3FDE / Engineer implementation
4. 4production use and evaluation signals

1. 1Account technical strategy
2. 2specialists and delivery teams
3. 3customer outcomes

1. 1Production failures and reusable patterns
2. 2Product / Research feedback
3. 3next implementation decision

1. 1People development and staffing
2. 2sustainable delivery

1. 1Training and workshops
2. 2adoption evidence

Read the first row as a question about handoff: what must be agreed before implementation, and what proves that deployment helped the customer? Read the feedback row as a separate output of field work. A working demo and a research-quality failure report are different artifacts. The last two rows help avoid presenting IC results as people management or training activity as production-code ownership.

Build your own evidence worksheet

This is an unexecuted preparation exercise, not a model answer or a claim about your experience. Pick one real, shareable project and your exact target vacancy. Fill these fields without using confidential customer data:

Worksheet field	What to record	What makes the evidence useful
Role and boundary	Vacancy ID; decisions/code you owned; others’ work	Separates your contribution from team results
Workflow	User, input, data/tool path, permission boundary, output	Explains why an implementation was needed
Baseline and measurement	Metric definition, period, denominator, comparison	Prevents unexplained “accuracy” or productivity claims
Trade-off	Alternative, rejected option, constraint, decision owner	Makes the judgment inspectable
Production evidence	Failure classification, observed signal, rollback/handoff	Connects shipping to stable customer use
Feedback	Reusable pattern or documented failure returned to product/research	Shows an additional field-work output
Unknown	Missing evidence and a way to verify it	Avoids filling gaps with confidence

If you cannot share an absolute customer number, explain a permitted range or normalized comparison and its limits. Do not manufacture a past achievement. A proposed design is acceptable preparation when clearly labelled proposed; it is not evidence of having shipped it.

What to confirm before applying

Ask about the particular job ID and round: format, duration, evaluation focus, AI/IDE/internet permissions and language. Tokyo FDE’s bilingual interviews and English résumé are explicit; those conditions were not confirmed in the DL posting and should not be copied onto it. Confirm ownership after launch, customer-site expectations, internal level and current opening status with the recruiter.

The official sources did not establish Tokyo FDE’s exact loop, take-home requirement, customer simulation, LeetCode use, coding language or framework. The one-week guide estimates are not SLAs. They also did not establish regional grade equivalence, Tokyo compensation, the DL’s people-management responsibility, the old title’s rename date, DeployCo’s relationship to the Tokyo employer or Tomoro’s closing. Keep these as questions; the practice articles will train reasoning from the documented work rather than invent answers to them.