What a forward deployed engineer actually is

AI Sales Engineering · v1.0 · Last verified 31 August 2026

A forward deployed engineer is an engineer who works inside the customer's environment rather than on the core product. They build and deploy the thing at the customer and carry the technical side of the relationship. At Anthropic the band is $280,000 to $320,000 OTE, checked 31 August 2026.

What is a forward deployed engineer?

A forward deployed engineer is an engineer who works inside the customer's environment rather than on the core product, building and deploying the thing at the customer. They carry the technical side of the relationship after signature as well as before it.

The title spread from Palantir into AI companies, which is why it turns up in frontier-lab postings and almost nowhere in traditional enterprise software. It sits between a solutions engineer and a software engineer. It isn't a support role, and reading it as one is the most common mistake people make about it.

What does a forward deployed engineer do?

Architecture and integration in the customer's environment, prototyping and building things that go to production, evaluation design, and debugging in situ. There's enough customer-facing work in it that discovery is part of the job rather than something handed to you.

Anthropic's Forward Deployed Engineer posting states the requirement directly:

"Production experience with LLMs including advanced prompt engineering, agent development, evaluation frameworks, and deployment at scale."

That's one sentence with four named skills in it, and evaluation frameworks is the one most candidates can't evidence.

What does a forward deployed engineer earn?

Anthropic's forward deployed engineer band is $280,000 to $320,000 OTE, checked 31 August 2026 from their live posting. For context, US sales engineer total compensation is about $200,000 at the median and $308,000 at the ninetieth percentile, and solution architect is $215,000 median and $345,000 at p90 (levels.fyi, August 2026).

Table: US total compensation percentiles against published frontier-lab bands. Sources: levels.fyi, August 2026, and Anthropic live job postings, checked 31 August 2026.

So the FDE band at a frontier lab sits at roughly the 75th to 90th percentile of the wider profession rather than in a different universe. These are US figures. Other markets differ, sometimes by a lot, and there's no equivalent published percentile data for most of them.

How do you become a forward deployed engineer?

The experience bar is lower than people assume. Anthropic's forward deployed engineer posting asks for 4+ years, and their Applied AI Architect role asks for 3+.

The named requirement across both is evaluation frameworks, so build one. For a real product, on realistic data, with the failure cases in it. Yours, not a template. That's the artefact two separate postings ask for by name and almost no candidate turns up with.

Then be findable, because for AI-titled technical roles 28% of hires come from sourcing against 37% from inbound (Ashby, data through Q2 2026, published 9 July 2026). Being sourced means someone looking for exactly your combination can find you, which makes publishing the artefact under your own name the highest-leverage thing you can do with a weekend.

Then outreach, to the hiring manager rather than the recruiter, with the artefact as the reason to be in touch. Then apply anyway, because inbound is still over half of all hiring generally.

Is it a good role?

Both sides, honestly.

Good: it pays at the top of the technical selling range. The work is real building rather than demo theatre, so what you learn compounds. Demand is growing fast and the supply of people who can do it is not.

Less good: it isn't remote at the frontier labs. Anthropic's postings specify a 25% minimum in-office expectation and are anchored to named cities. You carry customer pressure and engineering pressure at the same time, which is a specific kind of tiring that suits some people and wrecks others. And the title means different things at different companies, so read the posting rather than the title.

Forward deployed engineer versus solutions engineer

Table: How the two roles differ in timing, work and measurement. Source: AI Sales Engineering, from live postings and practitioner experience, August 2026.

The overlap matters as much as the difference. Both are customer-facing and technical, both do discovery, and the best people in either role look increasingly like the other.

Where to go next

If you want to know which parts of this you can already evidence, the four-axis capability model is the frame, and the diagnostic scores you against it. If you want the hiring mechanics in full, with the sources, read how AI sales engineering hiring actually works.

This comes from the same person who wrote those.

Sources, and what is not verified here

Last verified: 31 August 2026.

Fetched directly: Anthropic's live Forward Deployed Engineer and Applied AI Architect (Commercial) postings, including bands, experience requirements, quoted requirement text and the in-office expectation. levels.fyi for the US percentile figures, August 2026.

From published reports: the 37.1% inbound and 28% sourced split for AI-titled technical roles comes from Ashby, data through Q2 2026, published 9 July 2026. Ashby is one ATS vendor with a venture-backed customer skew, so treat it as the best available rather than the last word.

Not verified here: no non-US compensation data, because none is published at comparable quality. The description of what the role does day to day is practitioner judgement rather than a survey.

Corrections welcome, and they get credited.