Org breakdown / verified 2026-05-09
OpenAI Forward Deployed Engineer is one rung on a much deeper org chart.
Most pages on this topic scrape one job listing and stop. The actual OpenAI FDE function is a team called Model Deployment for Business (MDB) with eight distinct titles, a separate Gov variant, and a Hardware variant. Below is the verified breakdown of every open posting, what each title does in an engagement, and how the OpenAI version differs from the Anthropic one in ways that matter when you are sizing a deal.
Direct answer (verified 2026-05-09)
OpenAI's Forward Deployed Engineer is a senior IC on a team called Model Deployment for Business. The engineer embeds with one of OpenAI's most strategic customers and ships full-stack systems on top of frontier models, from first prototype through stable production. The SF and NYC listings pay $162K to $280K base plus equity. The OpenAI for Gov variant pays $145.8K to $280K base plus equity. The senior Forward Deployed Software Engineer track in SF pays $185K to $325K base plus equity. Hybrid 3 days a week in office, travel up to 50 percent.
Source: OpenAI careers search, FDE SF listing, Gov listing.
The team is called “Model Deployment for Business”, not Applied AI.
Every public OpenAI FDE listing in the commercial track names the same team on the JD: Model Deployment for Business, abbreviated MDB. That naming choice matters. Anthropic posts the same role family under Applied AI. The difference is more than branding. Applied AI signals an engineering function adjacent to research; Model Deployment for Business signals a function structured around accounts and revenue. The reporting lines, the measurement (production adoption, workflow impact, eval-driven feedback that changes product roadmaps), and the partner teams named on the JD (Product, Research, Partnerships, GRC, Security, and GTM) all line up with the second framing.
Two adjacent orgs use the same FDE shape but report elsewhere. The Gov track sits inside OpenAI for Gov with its own comp band and clearance requirement. A Hardware variant called Design Verification, Forward Deployed Engineering sits inside the Hardware org and covers semiconductor partner workflows. Same embed-with-customer pattern, different reporting chain, different deliverable surface.
Eight distinct titles inside the Model Deployment for Business org.
The shorthand “Forward Deployed Engineer” is one rung on a ladder that runs from sourcing to platform engineering management. Below is every distinct FDE-related title open under MDB on 2026-05-09. This is the org chart most write-ups skip.
Sourcer, Forward Deployed Engineering
- Where
- London, UK
- What it does
- Builds the candidate pipeline that everything below depends on.
Strategy and Operations, Forward Deployed Engineering (FDE)
- Where
- San Francisco
- What it does
- Designs how engagements are scoped, sequenced, and billed across the org.
Forward Deployed Engineer (FDE)
- Where
- 10 cities (US, EMEA, APAC)
- What it does
- Senior IC on customer engagements. Owns delivery from first prototype to stable production.
Forward Deployed Software Engineer (FDSE)
- Where
- San Francisco, Seattle, London, Tokyo
- What it does
- 7+ years full-stack experience. Pairs with FDEs and customer engineers, codes side by side.
Platform Engineer, Forward Deployed Engineering
- Where
- San Francisco, New York City
- What it does
- Builds the internal tooling and reusable abstractions FDEs ship with on every account.
Technical Deployment Lead, Forward Deployed Engineering
- Where
- San Francisco (incl. Platform), New York City
- What it does
- Leads multi-engineer deployments. Owns the delivery contract with the customer.
Manager, Forward Deployed Engineering
- Where
- London, Munich
- What it does
- People manager. Builds and runs a regional FDE pod against a regional customer book.
Platform Engineering Manager, Forward Deployed Engineering
- Where
- San Francisco
- What it does
- Manages the internal-platform engineers behind the FDE delivery surface.
The Sourcer and Strategy + Operations roles are not engineering roles, but they are public proof that OpenAI is staffing the org for scale rather than running it as a single ad hoc team.
Every open Forward Deployed Engineer IC posting, side by side.
Below is the IC FDE row of the org chart in detail: ten cities under Model Deployment for Business plus one Gov variant. Salary bands quoted are the public ranges on the JD on 2026-05-09. Where the listing does not post a salary, it is noted explicitly.
Forward Deployed Engineer (FDE), San Francisco
- Base salary
- $162K to $280K + equity
- Notes
- Hybrid 3 days a week. Travel up to 50 percent. 5+ years engineering experience.
Forward Deployed Engineer (FDE), New York City
- Base salary
- $162K to $280K + equity
- Notes
- Hybrid 3 days a week. Travel up to 50 percent.
Forward Deployed Engineer (FDE), Seattle
- Base salary
- Salary not posted publicly on the listing
- Notes
- Same role shape as SF and NYC.
Forward Deployed Engineer, Dublin, Ireland
- Base salary
- Not posted on the public listing
- Notes
- EMEA expansion track.
Forward Deployed Engineer, London, UK
- Base salary
- Not posted on the public listing
- Notes
- EMEA. Paired with a Manager and a Sourcer in the same city.
Forward Deployed Engineer, Munich, Germany
- Base salary
- Not posted on the public listing
- Notes
- EMEA. Paired with a Manager in the same city.
Forward Deployed Engineer, Paris, France
- Base salary
- Not posted on the public listing
- Notes
- EMEA.
Forward Deployed Engineer, Singapore
- Base salary
- Not posted on the public listing
- Notes
- APAC.
Forward Deployed Engineer, Sydney, Australia
- Base salary
- Not posted on the public listing
- Notes
- APAC.
Forward Deployed Engineer, Tokyo, Japan
- Base salary
- Not posted on the public listing
- Notes
- APAC. Posted alongside a Forward Deployed Software Engineer in the same city.
Forward Deployed Engineer, Gov, Washington DC, Seattle, San Francisco
- Base salary
- $145.8K to $280K + equity
- Notes
- Active TS/SCI clearance or equivalent required. Hybrid 3 days a week. Travel up to 50 percent.
What the SF FDE listing actually asks for.
People skim the FDE listing and miss the parts that matter at offer time: the 50 percent travel line, the hybrid 3-days a week in-office line, the explicit ask for production-grade code across frontend and backend, and the customer-facing experience requirement. Below is the public list as posted on the SF JD on 2026-05-09.
Responsibilities
- Own technical delivery across multiple deployments from first prototype to stable production.
- Build full-stack systems that deliver customer value and sharpen how we learn.
- Embed closely with customer teams, understand their needs, and guide adoption of what you build.
- Scope work, sequence delivery, and remove blockers early.
- Make trade-offs between scope, speed, and quality; adjust plans to protect delivery.
- Contribute directly in the code when progress or clarity depends on it.
- Codify working patterns into tools, playbooks, or building blocks that others can use.
- Share field feedback that helps Research and Product understand where the models succeed and where they can improve.
Requirements
- 5+ years of engineering or technical deployment experience that includes customer-facing work.
- Have scoped and delivered complex systems in fast-moving or ambiguous environments.
- Write and review production-grade code across frontend and backend using Python, JavaScript, or comparable stacks.
- Have built or deployed systems powered by LLMs or generative models and understand how model behavior affects product experience.
- Simplify complexity and make fast, sound decisions under pressure.
- Communicate clearly with engineers, product teams, and customer stakeholders.
What an OpenAI FDE actually lands in week 1 of a deployment.
The OpenAI JD describes the deliverable in delivery terms (full-stack systems, custom data pipelines, observable systems spanning infrastructure through applications), not artifact terms. That is a real difference from Anthropic's Applied AI listing, which names MCP servers, sub-agents, and agent skillson the JD. OpenAI's FDE leaves you with what you would expect a senior full-stack engineer to leave: a service, an eval surface, a small UI, and a deployment path.
The shape of week one looks something like this: a small service on the customer's infrastructure, calling OpenAI's API, with auth and a feedback loop to evals from day one. The exact stack varies by account; the JD names Python, JavaScript, and comparable stacks.
The customer keeps the file. The runtime calls OpenAI's API. The eval bus on every production call is the loop the JD names as “eval-driven feedback that changes product and model roadmaps.” The artifact is portable in the sense that the customer owns the code, but the model behind it is pinned to the vendor for the lifetime of the deployment.
Forward Deployed Engineer vs. Forward Deployed Software Engineer at OpenAI.
Same Model Deployment for Business team, same engagement, different role. The FDE owns the engagement; the FDSE is the senior full-stack engineer paired in to ship code on the customer's infrastructure. Both pull from the same intake. The salary and experience bands are different; the JD language is different.
| Feature | Forward Deployed Software Engineer | Forward Deployed Engineer |
|---|---|---|
| Years of experience asked for | 7+ years of professional full stack engineering experience. | 5+ years of engineering or technical deployment experience that includes customer-facing work. |
| Where the role sits in the engagement | Pairs with the FDE on the same account. Designs and ships the full-stack pieces side by side with the customer's engineers. | Owns the engagement end to end. Discovery, scoping, system design, build, production rollout, post-launch iteration. |
| What success is measured on | Solutions delivered, abstractions built that scale across accounts, knowledge captured back into the function. | Production adoption, measurable workflow impact, eval-driven feedback that changes product and model roadmaps. |
| Compensation band (SF listing) | $185K to $325K base + equity. | $162K to $280K base + equity. |
| Reports into | Model Deployment for Business team. Same org, different role on the engagement. | Model Deployment for Business team. |
| Public-listing language about background | Former founder, or early engineer at a startup who has built a product from scratch is a plus. | Customer-facing engineering background. Former technical founders fit the profile. |
The Gov variant: same role, different reporting line, lower comp floor.
The OpenAI for Gov FDE listing is the closest sibling to the commercial FDE, but it sits inside a different organization (OpenAI for Gov, not Model Deployment for Business). The differences that matter on the JD:
- Active TS/SCI clearance or equivalent.
- Familiar with cloud deployment models (Azure, AWS), Kubernetes, Terraform, and related infrastructure.
- Customer focus is defense, intelligence, and federal civilian agencies.
- 5+ years of engineering or technical deployment experience, ideally in customer-facing or government environments.
The compensation floor is lower ($145.8K vs. $162K) while the ceiling is the same ($280K). The team name changes from MDB to OpenAI for Gov. Customer focus is named explicitly: defense, intelligence, and federal civilian agencies. The infrastructure stack is named explicitly: Azure, AWS, Kubernetes, Terraform.
OpenAI FDE vs. Anthropic FDE, on the things that show up on the JD.
Two labs, two FDE programs, both senior IC roles embedded with strategic enterprise customers. The shape is similar; the wording on the public JD is different in ways that change how you should expect an engagement to run. Side-by-side comparison of what each lab actually posts.
| Feature | Anthropic FDE (Applied AI) | OpenAI FDE (MDB) |
|---|---|---|
| Team name on the JD | Applied AI. Single team across geographies. Federal Civilian variant on Applied AI. | Model Deployment for Business (the umbrella org). Plus the Gov track under OpenAI for Gov, plus a Hardware variant called Design Verification, Forward Deployed Engineering. |
| Number of distinct FDE-related titles open right now | Two: Forward Deployed Engineer, Manager, Forward Deployed Engineering. Plus Applied AI Architect as a sibling track. | Eight: Sourcer, Strategy and Operations, FDE, FDSE, Platform Engineer, Technical Deployment Lead, Manager, Platform Engineering Manager. |
| Cities the FDE IC role is currently posted in | Six US cities (Boston, Chicago, NYC, Seattle, SF, DC) plus Paris and London. Federal Civilian variant in SF, NYC, DC. | Ten: SF, NYC, Seattle, Dublin, London, Munich, Paris, Singapore, Sydney, Tokyo. Plus the Gov variant in DC, Seattle, SF. |
| What the JD says they ship | MCP servers, sub-agents, and agent skills, named explicitly on the JD. Artifact list shaped around Claude's tool-use surface. | Full-stack systems and custom data pipelines that create customer value. Scope is described in delivery terms (first prototype to stable production), not artifact terms. |
| Travel expectation on the JD | About 25 percent. | Up to 50 percent. |
| Office model | Office-attached, varies by listing. | Hybrid, 3 days a week in office. Relocation assistance offered. |
| Base comp band on the SF FDE listing | $200K to $300K USD on the Applied AI listing. | $162K to $280K + equity. |
Who actually gets an OpenAI FDE engagement.
Model Deployment for Business is positioned for OpenAI's most strategic customers. That phrase is doing a lot of work on the JD. In practice it has meant Fortune 500-scale enterprises with substantial existing or planned commercial commitments to OpenAI and a named internal sponsor who can match the FDE on technical ownership.
The bar to receive an embedded FDE rather than Solutions Engineering hours is roughly one of these three:
- A Fortune 500-scale enterprise with a real existing or planned commercial relationship with OpenAI.
- A regulated workflowin a vertical OpenAI is actively investing in (financial services, manufacturing, telecommunications all show up in OpenAI's public reporting on FDE outcomes).
- A federal or public-sector engagement, in which case the OpenAI for Gov variant applies.
If you are a $50M ARR Series B with a six-figure OpenAI bill and a senior MLE hiring gap, you are unlikely to get an embedded FDE. That is a feature of how the function is funded, not a comment on your project. The next section is for the segment that needs the same shape but does not qualify yet.
If you want the same artifacts, model-vendor neutral.
A Series A AI-native startup or a $2B to $20B enterprise with a senior MLE bottleneck rarely qualifies for an OpenAI FDE engagement. Here's what a model-vendor-neutral version of the same engagement looks like, and where it differs from the OpenAI role on the things that matter at handoff.
| Feature | OpenAI Forward Deployed Engineer | fde10x (vendor-neutral studio) |
|---|---|---|
| Customer eligibility | OpenAI's Model Deployment for Business team is for OpenAI's most strategic customers. In practice that's enterprises with substantial OpenAI commercial commitments and the matched internal ownership. | Series A AI-native startups ($8M to $25M raised) and $2B to $20B enterprises with a board AI mandate. We engage when the bottleneck is senior MLE capacity, not OpenAI inference commit size. |
| Model the agent runs on | OpenAI frontier models. The deliverables list and the eval-driven feedback loop are written against OpenAI's API and OpenAI's product surface. | Whatever the client picks: Anthropic, OpenAI, Bedrock, Vertex, Azure OpenAI, or open-weight. The same eval harness scores all of them. |
| What is in your repo at handoff | Production code on OpenAI models. The exact handoff artifact list is not on the public JD; engagement length and post-deployment ownership are not posted publicly. | Agent code, rubric.yaml, eval/cases.yaml, .github/workflows/pilot-gate.yml, runbook.md. No platform license, no vendor-attached runtime. You can swap models by changing one config line and re-running the rubric. |
| Engagement length | Engagement length not posted publicly. Role description names long-term partnership through deployment rollout and iteration. | Two to six weeks fixed-fee. Week 0 rubric, week 1 first PR, week 2 prototype in client staging, weeks 3 to 5 hardening, week 6 handoff with runbook and 90-minute transfer session. |
| Refund and exit | No public refund-and-exit clause. | If we miss the week 2 prototype rubric, billing pauses and you can exit at the calendar day 14 decision meeting. |
| Day-one access | FDE is embedded with the customer; specific day-one mechanics depend on the customer relationship and IT environment. | Named senior engineer in the client's GitHub, Slack, and standup on week 1. First PR within 7 days or billing pauses. |
If you have a substantial OpenAI commit and a strategic relationship with the company, an MDB FDE engagement is the right hire. This comparison is for the segment that does not qualify yet wants the same scope of full-stack delivery.
A few sentences on what actually happens in your repo.
Whether the engineer comes from OpenAI Model Deployment for Business or a vendor-neutral studio, the first week looks similar from outside the door. They land in your GitHub, your Slack, and your standup. They ship a first PR within seven days. The PR opens a folder for the agent, lays down a prompt or instructions file, and wires the runtime against whatever model the engagement is on.
The interesting question is what happens at week 6, not week 1. Whoever is in your repo, ask which file is the rubric, which file is the eval set, what runs the gate on every PR, and which of those files is yours when they leave. If the answer to all four is “yours, on your main branch,” the engagement is built to outlast the engineer. If any of those answers point at a vendor surface, you have bought a deployment, not a system.
Need an embedded FDE engagement and OpenAI MDB is not the right fit?
Tell us the agent you need shipped, the model your team picked, and the deadline. We scope a fixed-fee engagement on a free call.
Frequently asked questions
What is OpenAI's Forward Deployed Engineer role?
It is a senior IC on a team called Model Deployment for Business (MDB). The engineer leads complex end-to-end deployments of frontier models in production alongside OpenAI's most strategic customers. They own discovery, technical scoping, system design, build, and production rollout. As of 2026-05-09 the role is posted in ten cities (SF, NYC, Seattle, Dublin, London, Munich, Paris, Singapore, Sydney, Tokyo) plus a Gov variant in Washington DC, Seattle, and San Francisco. SF and NYC pay $162K to $280K base plus equity. The Gov variant pays $145.8K to $280K base plus equity. Hybrid 3 days a week. Travel up to 50 percent.
What is the difference between Forward Deployed Engineer and Forward Deployed Software Engineer at OpenAI?
Same team (Model Deployment for Business), different role on the same engagement. The FDE is the senior owner of the deployment: discovery, scoping, system design, production rollout, post-launch iteration. The FDSE is a 7+ year full-stack engineer who pairs with the FDE and the customer's own engineers, coding side by side on the customer's infrastructure. The FDSE listing in SF pays $185K to $325K base plus equity, against $162K to $280K for the FDE in the same city. The FDE listing asks for 5+ years and customer-facing experience; the FDSE listing asks for 7+ years of full-stack engineering with former-founder background as a plus.
How big is OpenAI's Forward Deployed Engineering org?
On 2026-05-09 the OpenAI careers search returns 22 active 'forward deployed engineering' postings under the Model Deployment for Business umbrella, in 13 cities, across 8 distinct titles: Sourcer, Strategy and Operations, FDE, FDSE, Platform Engineer, Technical Deployment Lead, Manager, and Platform Engineering Manager. Add the OpenAI for Gov variant (separate org but same FDE shape) and the Hardware track (Design Verification, Forward Deployed Engineering) and the public count is higher. The function grew from two engineers at the start of 2024 to 39 by year-end per public reporting, then continued to scale.
How much does OpenAI pay a Forward Deployed Engineer?
$162K to $280K base plus equity for the SF and NYC listings. $145.8K to $280K base plus equity for the OpenAI for Gov variant (DC, Seattle, SF). $185K to $325K base plus equity for the senior Forward Deployed Software Engineer track in SF. International listings (Dublin, London, Munich, Paris, Singapore, Sydney, Tokyo) do not post salary publicly on the careers page. Total comp including equity is not on the public JD; market commentary places top-of-band OpenAI offers significantly higher when equity is included.
What does an OpenAI FDE actually ship into a customer environment?
The JD names full-stack systems and custom data pipelines built against OpenAI's APIs, deployed inside the customer's infrastructure. Unlike the Anthropic Applied AI JD, OpenAI does not name MCP servers, sub-agents, or agent skills as the deliverable list. The OpenAI listing is written in delivery terms (first prototype to stable production, eval-driven feedback, observable systems spanning infrastructure through applications) rather than artifact terms. In practice the engagements have produced documented 20 to 50 percent efficiency improvements across financial services, manufacturing, and telecommunications deployments per OpenAI's own public reporting.
Can my company hire an OpenAI FDE for a project?
Model Deployment for Business engages with OpenAI's most strategic customers. In practice that is enterprises with substantial OpenAI commercial commitments and named internal sponsors who can match the FDE on technical ownership. If you are a Series A startup with a six-figure OpenAI bill and a senior MLE hiring gap, you are unlikely to receive an embedded FDE. The Solutions Engineering and Customer Success organizations cover smaller accounts. If you sit between the two and the bottleneck is senior MLE capacity, you have other options including model-vendor-neutral studios that ship a similar full-stack scope.
What is the OpenAI for Gov FDE variant?
Same role shape, different reporting line. The 'Forward Deployed Engineer, Gov' listing sits inside the OpenAI for Gov organization rather than Model Deployment for Business. It targets defense, intelligence, and federal civilian agencies. The compensation band is $145.8K to $280K base plus equity (vs. $162K to $280K for the commercial track). Active TS/SCI clearance or equivalent is required. Familiarity with Azure, AWS, Kubernetes, and Terraform is named on the JD. Posted in Washington DC, Seattle, and San Francisco.
What is Design Verification, Forward Deployed Engineering at OpenAI?
It is the Hardware-track variant of the FDE function. Posted under the Hardware org (not Model Deployment for Business), it covers chip-design and verification workflows for OpenAI's semiconductor partner deployments. It is a separate role from the customer-facing commercial FDE in cities like SF and NYC, but it shares the embedded-with-customer pattern. As of 2026-05-09 it is posted in two locations.
If I want a similar engagement but on a model my team picks, what does that look like?
A model-vendor-neutral version of the same shape. Senior engineer named in the scoping call lands in your repo on week 1. Week 2 prototype runs in your staging on an agreed rubric. Week 6 handoff includes the agent code, rubric.yaml, eval/cases.yaml, .github/workflows/pilot-gate.yml, and runbook.md, with the model provider chosen by you (OpenAI, Anthropic, Bedrock, Vertex, Azure OpenAI, or open-weight). The eval harness runs the same cases against any of them, so swapping models after handoff is a config change rather than a re-architecture. fde10x runs that engagement under a fixed fee with a refund-and-exit gate at calendar day 14.
Related guides
Forward Deployed Engineer at Anthropic: the Applied AI role and every open posting
Anthropic's Forward Deployed Engineer is a senior role on the Applied AI team. Same role family as OpenAI's MDB, different team name, different artifact list, different travel expectation. Side-by-side breakdown.
What is a forward deployed engineer? The role, the three flavors, and the four-file test
The role originated at Palantir under the codename Delta. By 2026 it has split into three contractually distinct flavors. The way to tell which one you hired is to ask which artifacts are in your repo when the engineer leaves.
AI forward deployed engineer: the four vendors selling the title in 2026
Model labs, big-firm consulting practices, platform vendors, and vendor-neutral studios. They look identical on the first call. The four-file test tells you which one is actually in your repo.
Comments (••)
Leave a comment to see what others are saying.Public and anonymous. No signup.