What is a forward-deployed engineer? The AI job explained in plain English

If you've seen the phrase "forward-deployed engineer" in tech headlines and wondered whether it involves the military — fair. It doesn't, but the metaphor is deliberate: an engineer deployed forward, out of the head office and into the customer's territory.

The short version

A forward-deployed engineer (FDE) is a software engineer who works inside a customer's business rather than back at the vendor. Their job is to make the vendor's technology actually work for that specific customer: understand the workflows, wire the product into the tools the team already uses, build the missing pieces, and stay long enough to make sure it sticks.

The role was pioneered at Palantir in the early 2010s. Palantir's customers — intelligence agencies, banks, manufacturers — could never fully specify what they needed up front, so Palantir stopped asking. It embedded engineers on site to watch how work actually happened and ship working software against reality instead of against a requirements document.

Why the role suddenly exploded

The AI boom made every company a Palantir customer, in one specific sense: they know the technology matters and they can't write down what they need.

A language model demo takes minutes. Making AI reliably useful inside a real business — with its messy data, legacy systems, compliance rules, and people — is the hard part. That gap between demo and deployment is exactly the gap FDEs exist to close, which is why AI companies now hire them aggressively (The Pragmatic Engineer has a good deep dive on the role and its demand).

What an FDE actually does all day

Strip the mystique and the job has four repeating motions:

1. Discovery. Sit with the team. Learn what the business actually does, where time goes, what breaks, what the constraints are.
2. Recommendation. Decide what to build or buy for this business — which tools, which approach, at what cost.
3. Build. Write and deploy the custom glue: integrations, automations, evaluations.
4. Stay current. Models, prices, and tools change monthly. What was the right answer in January often isn't in June. The FDE keeps re-deciding.

What does a forward-deployed engineer cost?

It's a well-paid job. Glassdoor's salary data puts typical base salaries well into six figures, and at the big AI labs total compensation runs far higher. Specialist agencies have started selling "FDE-as-a-service" engagements — embedded engineers on fixed-fee sprints — priced from roughly $8,000 for a couple of weeks upward.

Which is great if you're a funded startup or a bank.

What this means if you run a small business

You are not hiring an FDE. Neither is any dental clinic, restaurant, law firm, or five-person agency. The economics don't work at either end: you can't pay the salary, and the people doing this work are busy at enterprises.

But notice something about those four motions: only one of them — the custom build — strictly requires an engineer in your building. Discovery can be a well-designed questionnaire. Recommendation can be a maintained, budget-matched plan. Staying current can be a service that re-verifies prices and re-tests models so you don't have to.

That's the wager behind what we build at Dapols: productize the three motions software can honestly do, price them at small-business numbers ($29 for a plan, $99 for a custom setup), and be plain about the fourth — if you need custom code deployed inside your systems, that's our done-for-you service or a human hire, not an app.

The FDE model's core lesson applies at every business size: AI doesn't fail because the models are weak. It fails because nobody owned the gap between the technology and your workflows. Someone — or something — has to be forward-deployed.