'Forward Deployed Engineer' Is the Latest Hot Title in AI. 'Outcome Engineer' Is All You Need.

The Forward Deployed Engineer is the job title of the year. OpenAI stood up an FDE team and pays base salaries of $220,000 to $280,000 plus equity. Google Cloud lists FDE roles from $127k up to $265k and says it plans to hire hundreds. Anthropic runs the same job under the name Applied AI Architect. a16z called it “the hottest job in tech.” Levels.fyi now tracks it as its own salary track, next to backend and mobile.

Rewind eighteen months and the title of the year was “prompt engineer.” Before that badge cooled, it was “applied AI architect.” Now it is “forward deployed engineer.” Same conveyor belt, fresh sticker: the pay climbs, the label churns, and companies keep hiring the noun instead of the result.

So of course the question started landing in my inbox: “You do something like this, right? Should you be calling yourself a Forward Deployed Engineer?”

Sort of. And that “sort of” is the whole point of this post. The title is real and the demand is real. But the title describes what someone is. It says nothing about whether your number moves. Those are different purchases, and the gap between them is where most AI budgets go to die. If a title has to describe what actually matters, call it Outcome Engineer: the one badge that never goes stale, because it names the result, not the fashion.

Where the title comes from

Palantir coined it. The idea was borrowed from a forward-deployed soldier: stationed on the ground, close to the problem, ready to act. A Palantir FDE embeds with the customer, often onsite three or four days a week, learns how the business actually runs, then ships software that solves that specific problem.

Why did AI make this suddenly urgent? Not because the models got worse. Because the models got good and nothing changed. MIT’s 2025 State of AI in Business report found that 95% of enterprise generative AI pilots showed no measurable business impact. Accenture puts sustained, enterprise-wide impact at 32% of companies. The gap is not model quality. Models don’t deploy themselves. Someone has to sit inside the mess, wire the thing into the real stack, and make a number move. That someone is expensive and rare, so the title got hot.

Good so far. Here is the part the trend skips.

The classic FDE deploys a product. Someone else’s.

At Palantir, the FDE configures Palantir’s own platforms, Foundry and Gotham. At OpenAI, the FDE ships OpenAI’s models. At Anthropic, the Applied AI Architect deploys Claude into your stack. Palantir is explicit that this is not consulting: the FDE assembles a pre-built product to fit you, and the clever field work gets fed back so the product itself gets better. FDEs build many capabilities for one customer; the product team turns that into one capability for many customers.

That loop is the actual business model. The vendor eats the cost of embedding an expensive engineer because every deployment sells more product and sharpens it. The FDE is a spearhead for a platform.

Now strip the platform away. Copy the title without a product to deploy, and what is left? An engineer embedded in your company, building. That is a contractor. A good one, maybe. But you did not buy a spearhead, you bought a seat. This is the quiet trap: “hire an FDE” sounds like buying an outcome and often turns out to be buying staff augmentation with a fashionable name on the badge.

So “should I hire an FDE?” is the wrong question

Two reasons it misfires.

First, supply. The trait bundle that made Palantir FDEs work is genuinely rare: the social read to embed in someone else’s org without friction, the tolerance to sit in ambiguity, and the velocity to ship value before patience runs out. Ex-Palantir FDEs go on to found startups at a wild rate. You are hiring against OpenAI and Google for that exact profile, and they are paying a quarter-million-plus base. Everest Group’s read is blunter: the role is situational, not something every company needs.

Second, and this matters more: a title is an input. You can fill the seat and still not move the number. The 95% of pilots that went nowhere were not staffed by bad engineers. They were staffed against no defensible metric, in a demo environment, with no one owning what happened after the applause.

The question that actually pays: what number moves?

Strip the label and the useful part is a shape of work, not a job title. It looks like this:

• Pick one place AI moves a real number. Revenue up or cost down. Name it before any code.
• Ship it into your stack, behind your auth, with evals and observability so you can trust it.
• Hand it off. Runbooks, coaching, ownership on your team. The capability stays when I leave.

I keep saying “I” because this is the work I’ve been doing for a decade, long before it had a trending name. An in-house LLM pipeline that replaced a translation vendor and saved $337K a year, shipped in six weeks. AI sales agents at a unicorn-stage B2B SaaS that doubled outbound conversion. An employee referral engine at Beekeeper that now drives $1.5M a year and is the platform’s top revenue channel, running for 10M users.

None of those started with “hire someone forward-deployed.” They started with a number and a sponsor who owned it.

A buyer’s checklist

If you are about to hire, contract, or retain someone for AI deployment, four questions separate an outcome-owner from a seat-filler. Ask them before the title comes up.

1. What number does this move, and who signed off on it? If the answer is “explore AI” or “usage,” stop. There is no outcome to own yet.
2. Whose stack does it ship in, and behind whose auth? Real deployment lives in your systems. A demo lives in theirs.
3. Who owns it after the engineer leaves? If the honest answer is “no one,” you bought a dependency, not a capability.
4. Are they deploying a product they sell, or building what you actually need? Both are legitimate. They are not the same purchase, and the price and the incentives differ.

Outcome Engineer is all you need

Titles trend and cool. Prompt engineer, applied AI architect, forward deployed engineer: each was the hottest badge in AI for a season, then the next one arrived. Forward Deployed Engineer is a good one, because underneath the hype it points at something true: the gap between a capable model and a result is human, and someone has to close it in the field. But it will cool like the ones before it.

The one title that never goes stale is not on the conveyor belt. Call it Outcome Engineer: whoever owns the number, ships it in your stack, and leaves it moving after they are gone. That is not a fashion. It is the whole job.

You can call what I do forward-deployed engineering. I don’t lead with it, because the label is not what you are buying. You are buying a number that moves and stays moved. That is all you need. The rest is a sticker.

If you have a number in mind and a sponsor who owns it, book a 30-minute call. Bring the messy version of the problem.