Measured, not guessed
Forward Deployed Engineer: what the job postings actually ask for
I counted the words in 4,241 Forward Deployed Engineer and customer-embedded job postings, then re-counted from scratch against 515 live postings pulled from employer hiring APIs on the day of writing. The two counts disagree on 12 of 40 terms. That disagreement is on this page, because a resume built on the 28 terms both counts agree on is worth more than one built on a number that moves when you look again.
4,241postings in the corpus
515live postings re-verified
28terms both sources agree on
12terms they contradict
685postings in Europe
The list
Resume keywords that survived a second measurement
Live percentage is from 515 customer-embedded postings fetched directly from Greenhouse, Lever and Ashby APIs. Corpus percentage is from the earlier 4,241-posting collection. Bars are scaled against 100%, never against the largest value in the table, so a small number looks small.
| Term | Live | Frequency | Corpus | Delta |
|---|---|---|---|---|
| Python | 68.5% | 64.6% | +3.9 | |
| architecture | 66.6% | 71.5% | -4.9 | |
| travel | 57.3% | 42.7% | +14.6 | |
| stakeholder | 52.6% | 54.9% | -2.3 | |
| AWS | 45% | 44.3% | +0.7 | |
| Azure | 42.5% | 39.7% | +2.8 | |
| GCP | 41.2% | 42.6% | -1.4 | |
| deployment | 40.4% | 47% | -6.6 | |
| forward deployed | 40.2% | 51.9% | -11.7 | |
| scoping | 34.4% | 24.9% | +9.5 | |
| pre-sales | 33.8% | 22.2% | +11.6 | |
| implementation | 33% | 35% | -2 | |
| SQL | 28.9% | 30% | -1.1 | |
| TypeScript | 26.4% | 26.1% | +0.3 |
Python is the most frequent hard skill and the most stable, moving only
3.9 points between two independent samples.
If one term goes on the resume, it is that one.
Travel is the requirement candidates underestimate
Travel language appears in 57.3% of live postings, which puts it above most technical skills on the list. This role is customer-embedded by definition. A remote-first assumption is the wrong prior.
Cloud: no single vendor wins
- AWS 45%
- Azure 42.5%
- Google Cloud 41.2%
All three land within a few points of each other in both samples. A resume that reads as single-vendor is narrower than the market it is aimed at.
The 12 terms the two measurements contradict, and why they are not on the list
Each of these moved more than 15 points between samples. That is too large to explain by sample size, so none of them is safe to publish as a general claim. The honest reading is that the true value sits somewhere between the two columns and this data cannot pin it.
| Term | Live | Frequency | Corpus | Delta |
|---|---|---|---|---|
| agentic / AI agents | 8.3% | 42.2% | -33.9 | |
| API | 25.4% | 58.3% | -32.9 | |
| LLM | 13% | 42.9% | -29.9 | |
| enterprise | 39.8% | 67.4% | -27.6 | |
| SaaS | 9.5% | 30.6% | -21.1 | |
| embedded | 6% | 27% | -21 | |
| integration | 49.3% | 70% | -20.7 | |
| customer-facing | 63.7% | 44.4% | +19.3 | |
| ambiguity | 12% | 30% | -18 | |
| machine learning | 50.5% | 35.5% | +15 | |
| troubleshoot | 7.4% | 22.4% | -15 | |
| Kubernetes | 3.1% | 18.1% | -15 |
The likeliest cause is sample composition. Of 515 live postings, 344 are Databricks, whose customer-embedded roles are data and platform work rather than AI product work. That pulls AI vocabulary down and data vocabulary up. A larger, more evenly spread live sample is the fix, and it is not something I can assert my way out of.
Terms the first measurement never tested, frequent in the live sample
These were absent from the original keyword set, which is a gap in the list rather than a finding about the market. Worth carrying on a resume even though they have only one measurement behind them, which is stated here rather than hidden.
- production 43.9%
- end-to-end 35.1%
- executive 29.3%
- roadmap 27.6%
Low-frequency terms: what not to spend resume space on
Under 10% in the live sample. Real skills, but they are not what gets a Forward Deployed Engineer application read.
- observability 7.2%
- RAG 4.5%
- evals 2.5%
- LangChain 1.9%
- Docker 1.7%
- MCP 1.7%
- prompt engineering 1.6%
- Terraform 1.6%
- Go / Golang 0%
Geography
The European market is smaller than the job titles suggest
685 of 4,241 postings resolve to Europe, 16.2% of the corpus, covering 546 distinct openings across 175 companies. Only 140 are core Forward Deployed Engineer roles, from 47 companies.
| Country | Openings | Postings | Companies | Core FDE |
|---|---|---|---|---|
| United Kingdom | 237 | 312 | 118 | 59 |
| Germany | 79 | 95 | 57 | 15 |
| France | 46 | 55 | 29 | 10 |
| Spain | 41 | 51 | 26 | 10 |
| Netherlands | 37 | 49 | 28 | 9 |
| Ireland | 36 | 51 | 24 | 14 |
| Sweden | 28 | 36 | 19 | 6 |
| Italy | 24 | 26 | 13 | 4 |
| Poland | 16 | 16 | 10 | 3 |
| Europe (region stated, no country) | 15 | 20 | 12 | 1 |
| Switzerland | 12 | 16 | 13 | 4 |
| Portugal | 12 | 12 | 6 | 3 |
Counted on distinct openings, not postings, because employers file one
requisition per city and a single role advertised in eight cities is one job, not eight.
Remote does not mean you are eligible
Of 827 postings whose location field says remote, only 93 explicitly include Europe. 299 are United States or North America only, and 435 state no scope at all. If you are applying from Europe, the remote filter on a job board is telling you almost nothing.
9.6%
of European postings require a language other than English. German leads at 4.7%, then French at 2.2%. This is a floor: every source I could reach is an English-language board, so a posting written in German is not in this data at all.
5.7%
require a security clearance, concentrated in 14 companies. Mostly defence and government-adjacent work, and mostly a hard gate rather than a preference.
4.7%
mention work authorisation explicitly, and 1.5% state that sponsorship is not available. Low percentages, but a hard stop where they appear.
Career move
Moving from Solutions Architect to Forward Deployed Engineer
I measured the same terms across two title families in the European slice: Solutions Architect and adjacent pre-sales titles, versus the forward-deployed family. The gap is specific and it is not about cloud knowledge.
Transfers directly
Customer-facing work, stakeholder management, scoping, executive communication, architecture. These appear at similar or higher rates in Forward Deployed Engineer postings than in Solutions Architect postings. Nine years of pre-sales engineering is not a standing start.
Does not transfer
Quota and revenue-influenced framing. Pre-sales language appears far more often in Solutions Architect postings than in forward-deployed ones. A resume organised around influenced revenue is answering a question this role does not ask.
The actual gap
Shipped production code. Python is the top hard skill in the live sample at 68.5%, and production language appears in 43.9%. The gap is evidence of shipping, not knowledge of a cloud platform.
Counterintuitive result worth stating plainly: AWS is named in roughly the same
share of postings as Azure and Google Cloud, and cloud vendor depth is not what separates the two
title families. Deep single-vendor expertise is less differentiating here than it is in a Solutions
Architect role.
Provenance
How this was measured, and what it cannot tell you
Method, in order
- Collected 4,241 postings from named public sources: employer ATS endpoints (Greenhouse, Lever, Ashby, Workday), aggregators, and public RSS. Every row keeps its source URL.
- Classified on body text, not job title. A title filter on "forward deployed" misses 2,521 postings and 233 companies entirely, because the same job is advertised under a dozen different titles.
- Counted literal terms with word-boundary regexes, scrubbing decoys first. A bare match on "remote" catches "remote access VPN" and a payroll vendor called Remote.com, so those spans are blanked before counting.
- Re-fetched live postings from employer APIs and recounted with the identical regexes, so the two measurements differ only in the documents.
- Flagged every term that moved 15 points or more, and published the flags.
Reproduce: node .research/verify-keywords-live.mjs
What I can prove, and what I cannot
Provable. These postings exist, were fetched from named public endpoints, and term X appears in exactly M of them. Counted, not estimated, and re-runnable.
Not provable. That these percentages describe the world market. The denominator is "postings reachable on indexable, mostly English-language boards", and fourteen sources refused collection. National job boards such as StepStone, Welcome to the Jungle, InfoJobs and Pracuj are absent by construction, which means the non-English language requirement figure is a floor and the European counts are understated.
The measurement error that matters most. Frequency measures advertising language, not the job. Testing appears in under 3% of postings. That is not evidence that 97% of these teams skip testing. It is evidence that employers do not advertise it.
The live sample is skewed. 515 postings from 3 employers, of which 344 are one company. It is a falsification check, not a second corpus, and it is why 12 terms are marked contested rather than corrected.
Questions
Forward Deployed Engineer, answered from the data
What keywords should a Forward Deployed Engineer resume include?
Measured across 515 live customer-embedded postings, the terms that appear most often are Python (68.5%), architecture (66.6%), travel (57.3%), stakeholder (52.6%), AWS (45%), Azure (42.5%), GCP (41.2%), deployment (40.4%). Python is the single most frequent hard skill at 68.5% and it is the one term both independent measurements agree on most closely.
Is Forward Deployed Engineer the same as Solutions Architect?
No. Measured against the same posting bodies, pre-sales and quota language appears far more often in Solutions Architect postings than in Forward Deployed Engineer postings, while Python and production code appear far more often in Forward Deployed Engineer postings. The role is closer to a Solutions Architect who ships production code and owns the deployment after go-live.
How many Forward Deployed Engineer jobs are there in Europe?
Of 4,241 postings collected, 685 resolve to Europe (16.2%), covering 546 distinct openings across 175 companies. Only 140 are core Forward Deployed Engineer roles, from 47 companies. The United Kingdom accounts for the largest share by a wide margin.
Are Forward Deployed Engineer jobs remote?
Of 827 postings whose location field says remote, only 93 explicitly include Europe. 299 are United States or North America only, and 435 state no geographic scope at all. For a Europe-based applicant, remote does not mean eligible.
Do Forward Deployed Engineer jobs require travel?
Travel language appears in 57.3% of live customer-embedded postings, which makes it more common than most technical skills on the list. It is the requirement most often underestimated by candidates coming from a remote-first background.
Does a Forward Deployed Engineer need to know AWS?
AWS is named in 45% of live postings, Azure in 42.5% and Google Cloud in 41.2%. No single cloud dominates, so a resume that reads as single-vendor is narrower than the market. Multi-cloud framing matches the data better.
From the people doing the job
What the FDE leads at OpenAI, Ramp, Nominal and Dataland actually say
This page counts what job postings ask for. It cannot tell you what the work feels like, how teams decide where the line sits between an FDE and a consultant, or what each team screens for when hiring. So I indexed a panel of people who run the function: 34 questions, 69 answers, each one written in plain language and linked to the exact second in the recording so you can hear it in their own words.
Work with me
I build the thing this page measures
I am an AI Solutions Architect and Forward Deployed Engineer based in the Netherlands, with nine years at AWS deploying AI systems inside enterprise environments. I have spoken at AWS re:Invent and AWS Summits and published on the AWS Machine Learning and Architecture blogs.
Three things I do: deploy AI systems inside your customers' environments as an embedded engineer, review and rebuild Forward Deployed Engineer resumes against the data on this page, and train Solutions Architect teams making the move into forward-deployed work.