FDE Stands for Forward Deployed Engineer

US searches for the full title went from 206 a month in January 2023 to 17,084 in September 2026. That is 83x in 44 months, and almost none of it happened gradually.

Three fields use the same three letters. In firearms FDE is Flat Dark Earth, a paint color. In security it is full disk encryption. On the US results for the bare acronym, six of the seven organic results are the engineering role, so this page is the engineering one and does not try to be the other two.

Where the demand came from

The role is about fifteen years old and the search demand is fourteen months old. Palantir was running forward deployed engineers in the early 2010s and the volume did nothing about it for a decade. The line moves in July 2025, when US monthly searches for the full title went from 444 to 3,781 in a single month.

206

165

369

675

444

3,781

8,955

11,241

17,084

Jan
2023Dec
2023Jun
2024Jan
2025Jun
2025Jul
2025Nov
2025Mar
2026Sep
2026

Under that step is enterprise AI recreating Palantir's original delivery problem at industry scale. A frontier model is useful and useless at the same time out of the box, because the value only appears once it is wired into one company's data and workflows, and that work cannot be done from the vendor's office. So the vendors started shipping engineers instead of documentation. Andrew Ng published his definition of the AI FDE in May 2026 because people kept asking him about the career path after OpenAI and Anthropic started placing FDEs inside client organizations.

800-1000%growth in FDE postings
in a year, by source

$1BAWS investment in its own
FDE org, June 2026

22 of 311OpenAI open roles that were
FDE-shaped, June 2025

Two of those come from Andreessen Horowitz. Marc Andrusko puts postings up 800 to 1000% in a year, depending on the source, and Joe Schmidt counted 22 of the 311 open roles on OpenAI's careers page falling into the FDE and solutions category. The third number is AWS's own. In June 2026 it announced a $1 billion investment in AWS Forward Deployed Engineering to embed thousands of experts with customers. The acronym also has its own Wikipedia article now, which is the clearest sign a job title has stopped being one company's jargon.

What an FDE does that a normal engineer does not

The work is one full delivery cycle, owned end to end, run inside somebody else's company. Gergely Orosz reports that Palantir had more forward deployed engineers than conventional software engineers until around 2016, and quotes Palantir's own framing that the responsibilities look like a startup CTO's.

Discovery in situSit with the real workflow before deciding what to build

Build in their stackIntegration and custom logic inside the customer's systems

Production handoverInstrument it, prove it on real cases, leave it running

Saying noPush back when the request will not survive contact with the data

Roadmap return pathField gaps go back into the core product

One customerMany capabilities for them, not one feature for everyone

Ng's skill list is explicit that the hard parts are not all technical. An FDE has to understand the client's needs, prioritize the work, explain the technology, and push back when the ask is unrealistic. An engineer who cannot hold that conversation will build the wrong thing beautifully.

Two adjacent acronyms travel with it. FDSE is forward deployed software engineer, Palantir's original title, still the standard phrasing where the deployed artifact is conventional software. AI FDE is the 2025 and 2026 version, where the artifact is a model, an agentic workflow or an eval suite. FDE covers all of them, and the posting tells you more than the title does.

The FDE nobody has written down is the go-to-market one

I pulled the US results for the bare acronym before writing this page, and pulled them again on 6 September 2026. The two sets barely overlap. The 28 August crawl held Wikipedia at 2, a personal site at barry.ooo at 4, Invisible Technologies at 5, futureventures.ca at 6, fde.academy at 9 and Rocketlane at 10. Nine days later the top ten read IBM at 2, a Greg Isenberg post on X at 4, a different fde.academy article at 5, Reddit at 6, a full disk encryption explainer at 8, and two LinkedIn posts at 9 and 10. One domain held its place.

Two things follow. Weak domains rank here: barry.ooo is DR 21 and fde.academy is DR 26, and one of the two is on page one in both pulls. And no page on this result is protected by incumbency, because the set rewrites itself faster than anyone can build a position on it.

The page to beat is barry.ooo. Barry McCardel spent nearly five years in a forward deployed role at Palantir and now runs Hex, and his essay is the best writing on this search result by a wide margin. His argument is that most companies copying the model are, in his phrase, doing sparkling sales engineering, and he earns it twice over: on the economics, where a deployment is R&D rather than cost of goods and is supposed to lose money, and on the organizational condition, where field teams have to be trusted to decide. I am not going to pretend that page is weak.

It leaves two things open. It never says what FDE stands for, because it is written for people who already know. And it is addressed to founders deciding whether to build an FDE organization, not to a company deciding whether to hire one; it closes by explaining that Hex does not employ FDEs. I also read the five other substantive pages across both pulls, Invisible Technologies, futureventures.ca, fde.academy, Rocketlane and IBM. Every one of them puts the deployment target in the product or data organization. Not one applies the model to a revenue team. That is the gap this page is written into.

The reason the model carries over has little to do with AI. It carries over because domain context cannot be shipped remotely, and that condition holds anywhere software only pays off once somebody understands the business. Go-to-market is the clearest remaining example.

Why revenue teams are the next deployment target

GTM tooling underdelivers for a reason that has nothing to do with the tools. Clay, Unify, Common Room, Cargo and Octave all work as advertised. What fails is the assembly: the signal logic that only makes sense against one company's ICP, the enrichment tuned to one data reality, the routing rules that encode one team's actual process, the CRM debt nobody outside the building can see. Buying another tool does not close a deployment gap.

The signals are already in the record, uncommented. PostHog's explainer notes in passing that FDE roles now range from product development to sales to RevOps, then drops the thread. Andrusko describes putting engineers on planes to land seven-figure deals, which is a revenue motion written in engineering language. And Clay, which coined the title GTM engineer in 2023, defines that role at the intersection of commercial thinking and technical building. Point that person at one company for a fixed period with a mandate to ship, and you have described a forward deployed GTM engineer without using the phrase.

One engagement, and what it did

I have run this model. Over the past year I led the go-to-market transformation at Backbase, the banking platform that more than 120 banks run on, and the shape of the engagement was the one this page describes. I embedded with the revenue organization rather than the product organization, built against their accounts and their data rather than a reference implementation, and the exit condition was a system the team owns and runs without me. That last part is the production handover, and it is what separates this from advisory work.

The numbers are the client's rather than mine. Backbase CMO Tim Rutten published them himself in Topline on 19 May 2026. Account coverage per ABM marketer up 10x. New and qualified pipeline up more than 85% year on year, average deal size up more than 60% year on year, on a marketing budget optimized by 25%. He also sets out what the build replaced: the ABM platform, the integration stack and the reporting stack were all absorbed into something the team built itself.

That last sentence is the deployment result, not a procurement saving. Three bought systems were retired because an engineer sat inside the revenue organization long enough to learn what those systems were actually being asked to do, then built that against their data. A vendor could not have shipped it from outside, which is the entire argument for putting the engineer in the building.

One engagement is one engagement, and I am not going to dress it up as a pattern. A single named case cannot tell you the model generalizes. What it can tell you is that the role defined above is one somebody has actually filled, with an operator willing to put his own name and his own numbers to the outcome.

If you searched the acronym because you are hiring

You need an FDE when your tools already work, nothing is wired together, and the wiring has to ship this quarter. The demand data answers the supply question before you ask it. Postings are up 800 to 1000% in a year, AWS committed a billion dollars to an FDE org, and OpenAI is hiring the title across eight cities, on a count of its job board taken 8 September 2026. Budget is rarely the blocker. People who have done the job are, and most of them are already deployed somewhere.

• Is the blocker a tool or an assembly? If the stack is fine and nothing is connected, another tool will not help.
• Can the person write code that stays? An advisor who leaves a deck has not deployed anything.
• Who owns it after the engagement? The handover is the deliverable, not an afterthought.
• Is the metric a system or a slide? Production adoption for a product FDE, pipeline for a GTM one. Pick before you start.

The internal version of this hire is the growth engineer, and the org design that receives either one is the part most companies get wrong first.