Ten sources,
each wrong about something
Most writing about forward deployed engineering is recruiting copy or speculation. A small number of pieces are genuinely load-bearing: first-hand accounts by people who did the job, primary documents worth reading as documents, and market data. Here is the shelf, in reading order, with what to take from each — and what each one is silent about, because that matters as much.
Start here: what the job feels like
Two first-hand accounts. Read these before any job description, or you will read the job descriptions wrong.
Nabeel Qureshi — “Reflections on Palantir”14
Take: the texture of forward-deployed work at the company that invented the title — the Airbus engagement, the culture, what it feels like to be dropped into an organisation and expected to produce. It is the best answer available to “what is this actually like?” and it is unusually honest about the parts that were bad.
Silent about: the modern AI-native version of the role, and compensation. It describes one company, one era, one person’s experience — which is its strength as a document and its limit as a guide.
PostHog — “WTF is a forward deployed engineer?”15
Take: the same role at a product company rather than a platform company, written first-party and plainly. It is the clearest available account of the travel reality, which most sources skip and which is the single most common reason people leave the role.
Silent about: enterprise-scale deployment concerns — security review, air-gapped environments, the commercial layer. A smaller-company view, and the more relatable one.
Baseten — “What I learned as an FDE at an AI startup”31
Take: the AI-native version, at the stage where nobody has patterns yet. Useful precisely because it is not from a mature FDE organisation — it describes the improvisation that a polished job description hides.
Silent about: what scales. One engineer’s tour at one startup; read it for the shape of the days, not the method.
Job descriptions describe the role a company wants to hire. First-hand accounts describe the role as lived, including the parts nobody would put in a posting: the travel, the loneliness of being the only engineer who understands the system, the weeks where you are a project manager with a laptop. Reading the postings first calibrates you to the marketing.
Primary documents: read the job descriptions as documents
Not to apply — to see what companies are committing to in writing.
| Source | What to take | What it is silent about |
|---|---|---|
| OpenAI — FDE job description9 | The phrase “full arc of a deployment”. It is the clearest single statement of the builder variant, and it is what to hold a company to in an interview. | How many customers at once, and whether any of the comp is variable. Both are askable. |
| Anthropic — FDE, Applied AI10 | A published 40/30/30 split and MCP servers, sub-agents and agent skills named as deliverables. The 30% that feeds back into the product is the part that distinguishes this from consulting. | What happens when the 40% expands to fill the quarter, which is the normal failure mode of that split. |
| Anthropic — FDE, Federal Civilian29 | Air-gapped and impact-level deployment stated as requirements. Read it if you want to understand the regulated end of this work — it is a different job. | The personal cost of clearance work: travel constraints, portfolio invisibility, on-site requirements. |
| Palantir — “Dev versus Delta”13 | The two-role split — FDSE and Deployment Strategist — that most companies copied the title of without the structure. It explains why one word now means two jobs. | That almost nobody else adopted the pairing, so a single FDE elsewhere absorbs both accountabilities. |
Read them for what a company is committing to rather than what it is asking for. “Owns the full arc of a deployment” is a promise about scope; a published time split is a promise about how your quarter is allocated. Both are legitimate things to ask about in an interview, and asking signals that you read carefully — which is the trait the job is about.
The commercial argument, and the market data
a16z — services-led growth8
Take: the reason the role exists, stated as a trade — margin for defensibility. If you understand this essay you can predict most decisions your employer will make about your team.
Silent about: the failure mode it walks into. It argues for the trade and does not dwell on organisations that pay the margin and never receive the moat. Read alongside our breakdown.
MIT NANDA — the 95% figure17
Take: the demand-side explanation. Capability is available and outcomes are not, which is a deployment gap rather than a model gap.
Silent about: its own methodology — the figure circulates through press coverage rather than a broadly available paper, so quote it as approximate. Our breakdown handles the caveats.
Bloomberry — 1,000 FDE jobs analysed1
Take: that one title covers three different jobs, and roughly a third of postings are pre-sales-shaped. This single fact will save you from accepting the wrong role.
Silent about: filled roles. Posting counts measure demand signals, and title inflation is real in a hot market. Breakdown here.
First Round Review — hiring an FDE11
Take: the hiring side, which tells you what is actually being assessed. It puts the engineering bar at staff level plus a temperament most strong engineers have not been asked to develop — useful because it names the second half explicitly.
Silent about: what the role does to a career over five years, which is the question candidates should be asking.
The Pragmatic Engineer — forward deployed engineers12
Take: the industry-wide framing, written for engineers rather than for investors or recruiters. The best single orientation piece if you read only one.
Silent about: the practitioner detail — evals, identity, the security review — which is where the work actually lives.
Where the role is going
Latent Space — forward deployed engineers at the AI Engineer World’s Fair30
Take: the 2026 shift some teams describe — from writing feature code to writing the systems that write feature code, with engineers supervising agent harnesses. If that direction holds, it changes what the deliverable is, and the agent harness design is our attempt at what it implies.
Silent about: whether it generalises. One company’s direction of travel, described at a conference, is a hypothesis rather than a trend.
Palantir blog — a day in the life33
Take: the concrete shape of a week at the origin company, which is a useful anchor against which to read every newer description.
Silent about: everything after 2020-era Palantir. Read it as a historical baseline, not a current spec.
There is very little good writing on the parts of the job that consume the most time: enterprise identity, the security review, the commercial layer, and evaluation for client work. That is not an accident — those topics are unglamorous, company-specific and rarely written up by the people doing them. It is also why most of this site exists. If you know of a genuinely good first-hand piece on any of them, it belongs here.
The reading is the cheap part
Every source above describes the same underlying skill: sitting with ambiguity in front of someone who is paying, and producing a next step rather than a caveat. That is practisable.
Quick answers
What should I read first to understand forward deployed engineering?
First-hand accounts before job descriptions. Nabeel Qureshi’s "Reflections on Palantir" for the texture of the work at the company that invented the title, PostHog’s first-party account for the travel reality that most sources skip, and Baseten’s account for the AI-native version at a stage where nobody has patterns yet. Reading postings first calibrates you to recruiting language rather than to the job.
Why read FDE job descriptions as documents?
Because they show what a company is committing to in writing, not just what it wants. "Owns the full arc of a deployment" is a promise about scope; a published time split such as Anthropic’s 40/30/30 is a promise about how your quarter is allocated. Both are legitimate to ask about in an interview, and asking demonstrates the careful reading the job itself requires.
What explains why demand for the role grew?
Two sources together. a16z’s services-led-growth argument gives the supply-side logic — deliberately trade gross margin for deployment depth because depth is defensible. The widely reported finding that roughly 95% of enterprise AI pilots showed no measurable P&L impact gives the demand-side logic: capability is available and outcomes are not, which is a deployment gap rather than a model gap.
What is missing from the available writing about this role?
Almost everything about the parts that consume the most time: enterprise identity and SSO, the vendor security review, the commercial layer of SOWs and pricing, and evaluation for client work. Those topics are unglamorous, company-specific and rarely written up by the people doing them, which is why most practitioner material has to be assembled rather than found.
Where is the role heading?
One direction described at the 2026 AI Engineer World’s Fair is a shift from writing feature code to writing the systems that write feature code — engineers supervising agent harnesses rather than authoring every change. Treat it as a hypothesis rather than an established trend: it is one company’s direction of travel, described at a conference, and it has not yet been shown to generalise.
Sources
Every number on this page traces to one of these. Where a figure is self-reported or crowd-sourced rather than first-party, it is labelled inline.
Cited on this page
- Bloomberry — “I analyzed 1,000 forward deployed engineer jobs” — posting growth, builder/pre-sales/internal split. bloomberry.com/blog/i-analyzed-1000-forward-deployed-engineer-jobs-what-i-learned/
- a16z — Services-Led Growth — the margin-for-moat thesis behind FDE hiring. a16z.com/services-led-growth/
- OpenAI — Forward Deployed Engineer job description — “full arc of a deployment”. openai.com/careers/forward-deployed-engineer-(fde)-nyc-new-york-city/
- Anthropic — Forward Deployed Engineer, Applied AI (JD) — 40/30/30 split; MCP servers, sub-agents and agent skills as deliverables. jobs.menlovc.com/companies/anthropic/jobs/69674588-forward-deployed-engineer-applied-ai
- First Round Review — “So you want to hire a forward deployed engineer” — the staff-engineer bar. review.firstround.com/so-you-want-to-hire-a-forward-deployed-engineer/
- The Pragmatic Engineer — Forward deployed engineers — newsletter.pragmaticengineer.com/p/forward-deployed-engineers
- Palantir blog — “Dev versus Delta” — FDSE and Deployment Strategist as distinct forward-deployed roles. blog.palantir.com/
- Nabeel S. Qureshi — “Reflections on Palantir” — first-hand account of forward-deployed work, including the Airbus engagement. nabeelqu.co/reflections-on-palantir
- PostHog — “WTF is a forward deployed engineer?” — first-party account of the role at a product company, including travel reality. posthog.com/blog/forward-deployed-engineer
- MIT NANDA “State of AI in Business” — 95% of pilots with no measurable P&L impact — widely reported; figure cited via press coverage rather than a public PDF. techcrunch.com/2026/07/30/forward-deployed-engineers-are-the-ai-industrys-latest-talent-obsession/
- Anthropic — Forward Deployed Engineer, Federal Civilian (JD) — air-gapped and IL-level deployment requirements. jobs.menlovc.com/companies/anthropic/jobs/76278352-forward-deployed-engineer-federal-civilian
- Latent Space — Forward deployed engineers (AI Engineer World’s Fair) — Ramp’s shift toward engineers supervising agent harnesses. www.latent.space/p/forward-deployed-engineers-aiewf
- Baseten — “What I learned as a forward deployed engineer at an AI startup” — www.baseten.co/blog/what-i-learned-as-a-forward-deployed-engineer-working-at-an-ai-startup/
- Palantir blog — a day in the life of an FDSE / Deployment Strategist — blog.palantir.com/
Explore the topic
See this alongside everything else on the same subject — handbooks, system designs, challenges and tools, in one place.
More Handbooks
- The Prompting HandbookA friendly, hands-on field guide for everyday humans — learn the CRISP framework, spot bad prompts, practice with real recipes, play a drag-and-drop game, and test yourself with a quiz. No code required.Read →
- The Agentic AI Interview HandbookTwenty topics every senior AI engineer should be able to reason about live — from eval pipelines to reliability patterns for generative systems.Read →
- The Senior AI Engineer Interview Handbook60 questions across architecture, production incidents, agentic systems, RAG, evals, cost, safety, and leadership — what staff-level AI interviewers actually probe for.Read →
- 50 Angular Interview QuestionsA visual handbook covering components, change detection, RxJS, signals, routing, forms, performance, and testing — what interviewers actually probe for in senior Angular roles.Read →
- 50 Python Interview QuestionsFundamentals to advanced: data structures, OOP, iterators & generators, the GIL, asyncio, memory, testing, and the standard library — a visual walk through everything a Python interview touches.Read →
- 51 LLM Evals Interview QuestionsGolden sets, LLM-as-judge, regression testing, offline vs online evals, RAG evals, agent evals, red-teaming, and observability — demystified for interviews and production.Read →
Explore more from Vibe Engines
Get the next one in your inbox.
New handbooks, system-design walkthroughs, and tools — straight to your inbox. No spam, unsubscribe anytime.