Whether it’s a quick question or a detailed brief — we’d love to hear about it.
AI explained
What is a forward deployed engineer, and should you hire one?
A forward deployed engineer is a software engineer who works inside a customer’s business instead of from their own company’s office. They sit with the people who own a process, build software against the customer’s real data and systems, deploy it there, and stay until it runs.
AI
Knowledge-Base
Overview
Where the role came from, what the job involves, what it pays in the US and the UK, and the question most guides skip: whether your business should hire one, or bring in a team that already works this way.
The title comes from Palantir. The way of working is now everywhere AI meets a real business, including how we build dedicated AI development teams.
Where the role came from
Palantir created the role in the early 2010s, with the internal name “Delta”. Its government customers had data spread across systems nobody could join up, and no engineers of their own to do it, so Palantir sent its engineers to sit inside each agency and make the software work there. Until around 2016, it had more Deltas than conventional software engineers.
“Forward deployed” is borrowed from the military, where a forward-deployed unit is stationed at the front rather than at headquarters. Palantir’s own distinction is the clearest: a normal developer works on one capability for many customers; a forward deployed engineer works on many capabilities for one customer.
Why everyone is hiring them in 2026
Because AI is easy to demonstrate and hard to put into production. A model that performs in a demo still has to be connected to a bank’s loan system or a lender’s document store, with that company’s data, security rules and older systems. MIT research in 2025 found that around 95% of enterprise AI pilots fail, largely because company data is siloed and hard to connect.
Someone has to close that gap by hand, and that is what the role is for. Forward deployed engineer job postings rose by more than 800% between January and September 2025, according to Financial Times data, and OpenAI, Anthropic, Microsoft and Salesforce all now hire for it. Andreessen Horowitz describes the strategy as trading margin for moat: the hands-on work costs the vendor profit in the short term and earns it customers who can’t easily leave.
What a forward deployed engineer actually does
Day to day, the job looks less like a software role and more like being the customer’s own engineer:
- Joins the customer’s standups and works to the customer’s priorities, not their employer’s roadmap
- Writes production code in the customer’s environment, against the customer’s data
- Turns half-formed requests from non-technical managers into working specifications, often live in the room
- Builds the integrations and data pipelines that make a model usable in that specific business
- Fixes what breaks, on the customer’s deadline
- Feeds what they learn back to the core product team, which is where much of the value sits
One engineer who has done the job at Palantir and Rippling describes it as three hats at once: consultant, product manager and software engineer.
Forward deployed engineer vs software engineer vs solutions engineer
The titles get mixed up, so here is the difference on the things that matter.
| Software engineer | Solutions engineer | Forward deployed engineer | |
|---|---|---|---|
| Owns production code | Yes, in their own product | Rarely | Yes, in the customer’s systems |
| Faces the customer | No | Yes, mostly before the sale | Yes, closely, after the sale |
| Measured on | Uptime, adoption, code quality | Deals won | Go-live dates hit, customer outcomes delivered |
| Builds for | Many customers | One sale at a time | One customer, many capabilities |
A solutions engineer shows what the product could do. A forward deployed engineer stays and makes it do it.
What a forward deployed engineer is paid
Recruiter data from Paraform puts US base salary between $150,000 and $217,000, with a midpoint of $183,000. A separate review of 1,000 job postings by Bloomberry found a median base of $173,816. At the AI labs, total pay for mid-to-senior roles runs from $350,000 to $550,000 once equity is counted. In the UK, QA puts the typical salary at around £85,000.
Those figures are the cost before the hiring itself. The role needs a strong engineer who is also good with people, and the AI labs are competing for the same candidates.
Should you hire one?
There are two honest answers, depending on what kind of business you are.
If you sell software, and deals are stalling because customers can’t get your product live inside their systems, then yes. That is the problem the role was invented for, and the trigger is usually one specific contract. Go in knowing the trade: the work lowers your margin, and if the team drifts into building one-off projects instead of improving the product, you have become a services company without deciding to.
If you run a business that wants AI in its own operations, you don’t need the job title. You need the way of working: an engineer who sits with your people, builds against your real data and systems, and is measured on whether the process got faster and cheaper. You can hire for that at the salaries above, in a market where postings rose more than 800% in nine months. Or you can bring in a team that already works that way.
That is how our teams work. When a US mortgage lender came to us, three developers started on one document review that took its credentialed staff 45 to 50 minutes a file. The team built the tools into the lender’s loan origination system, with checklists the lender’s own experts can edit. The first tool cut the review to 15 minutes and saved more than $40,000 a month. Within nine months, more than 12 tools were in production, saving more than $175,000 a month across three core workflows. Read the case study.
Every engineer on those teams works to the same playbooks, a named developer owns every output, and each team member costs from $4,500 a month rather than a US salary, because the team works from the Philippines under UK leadership.
How to tell a real forward deployed engineer from a contractor with a new title
The title is fashionable, so some of what is sold under it is ordinary outsourcing. Whoever you hire or bring in, check for these:
- They work in your repository and your tools, not in a separate environment you only see in demos
- They join your standups and talk directly to the people who do the work
- One named person is accountable for each thing that ships
- They measure the process before they touch it, so the improvement is a number
- They leave runbooks and documentation your own people can run
- What they learn goes back into how the next build is done, not just this one
If someone can’t say yes to all six, you are buying a contractor, whatever the title says.
Questions people ask
What does FDE stand for?
Forward deployed engineer. Some companies use forward deployed software engineer, implementation engineer or customer engineer for the same job.
Is a forward deployed engineer a consultant?
No. A consultant recommends and leaves. A forward deployed engineer writes production code in your systems and stays until it runs.
Is a forward deployed engineer the same as a contractor?
Not if the role is done properly. A contractor delivers to a brief from wherever they sit. A forward deployed engineer works inside your business, owns the outcome and feeds what they learn back into the way the work is done.
Can a small or mid-sized business get this without hiring?
Yes. The way of working doesn’t depend on the title. A dedicated team that works inside your business does the same job, starting with one engineer, and you measure it the same way: the new process against the old one.
How quickly does it make a difference?
With our teams, a proof of concept on one process takes three days, and most teams are operational within weeks. What decides how fast it pays is starting with one process that has a clear cost, not with a strategy. The lender above started with a single document review, and its first tool saved more than $40,000 a month.
Want the way of working without the hire?
See how we build dedicated AI development teams, or tell us which process costs you the most and we’ll say honestly whether it would pay back.