Hiring an FDE firm or your own forward deployed engineers

A company that wants engineers working inside its own teams has two ways to get them. It can hire forward deployed engineers as its own employees, or it can contract a firm that supplies them. This page compares the two, using only public figures, each with its source. The guide to forward deployed engineering explains the role itself, and its page on how to hire a forward deployed engineer covers recruiting in detail.

What the two paths mean

Your own forward deployed engineers are employees. You find them, interview them, pay them a salary and benefits, and manage them. They stay when the first project ends, and what they learn stays with your company.

An FDE firm employs the engineers, or in some cases works with contractors, and places them inside your team under a contract. You pay the firm. The firm finds the people, replaces anyone who leaves, and brings its own method. When the contract ends, the engineers go back to the firm.

A third path, staff augmentation, adds engineers who take tasks from your own leads. The FDE guide compares it in forward deployed engineer vs staff augmentation.

What public pay data shows

Two public sources give pay figures for the role.

Source	What it covers	Figure
Levels.fyi [1]	Forward deployed engineers at Palantir in the United States, from 226 salary reports that users submitted, updated 15 October 2026	$201K to $590K a year; median $233K
Tenex careers page [2]	Pay for Tenex's own Forward Deployed Engineer role, on site in New York City, read 20 October 2026	$224K to $394K

Both figures are pay to an engineer, and a firm's price to a client is a separate number that no firm in this guide publishes. The Levels.fyi range covers one employer, and its figures come from people who chose to report them. The Tenex range covers one role at one firm in one city.

An employer pays more than the salary. It also pays for benefits, recruiting, equipment, managers' time and the weeks before a new hire is fully working. A firm's price has to cover the same costs for its own staff, plus its offices, its sales work, the time its engineers spend between projects, and a profit. So a firm's rate for an engineer will be higher than that engineer's pay. Whether the total cost is higher depends on how long you need the work and how fast you could hire.

Why hiring your own can take time

Henley Wing Chiu of Bloomberry, using job data from Revealera, found that postings for forward deployed engineers from January to October 2025 were 1,165 percent higher than in the same months of 2024. He found that October 2025 had the most FDE postings of any month in the data, and that 58 percent of the roles were at companies with 11 to 200 employees [3]. More postings mean more employers looking for the same people. Bloomberry describes most of those employers as growing startups, and a company hiring its first FDE competes with them for the same candidates.

Joe Schmidt of a16z wrote in June 2025 that, at the best companies that sell complex AI applications, the person who makes the product work for each customer is often a professional services employee, "sometimes rebranded as a forward deployed engineer" [4]. He counted 22 of the 311 open roles on OpenAI's careers page in those categories at the time. His essay argues that software vendors accept lower margins, meaning a smaller share of each sale left as profit, to pay for this work. It is written for software vendors, so use it as evidence of the demand, and of why those vendors spend on the role.

What a firm adds

• People sooner. A firm can place engineers it already employs. Your own search has to begin by finding candidates.
• A replacement when someone leaves. The firm finds the next person, and the contract can say how fast.
• A method. Each firm brings its own way of working. Reveneau, by its own account, has AI write all of its code and requires every change to pass a large eval suite, a set of automated tests written from the specification, before release.
• Skills you need for one stage only. A firm can add a data engineer for a month and then remove the role.

What a firm takes away

• The knowledge, unless it is handed over. When the engineers leave, what they learned about your systems leaves with them. A written handover plan in the contract is the fix.
• A share of the price as margin. You pay for the firm's own costs and profit.
• Some control over who does the work. The firm picks and moves its people. TechCrunch reported in 2024 that Tribe AI builds client projects with a network of over 500 contractors [12], which is one reason to ask who exactly will work on your project.

What one MIT report found about outside partners

MIT's Project NANDA report, The GenAI Divide, published in July 2025, found that in its sample, customized AI tools that can learn from use, built with external partners, reached deployment, meaning they were put into use, 67 percent of the time. Tools built inside the company did so 33 percent of the time. The report marks both figures as approximate [5]. It also says the figures come from what companies reported about themselves, may not account for every other cause, and are preliminary. The site that hosts the copy read for this guide summarizes its method as 52 interviews with executives, 153 surveys and more than 300 reviews of AI projects.

The report compares outside partners in general with in-house builds and gives no separate figure for FDE firms. A single report should not decide a large purchase. It is the only public study found for this guide that compares outside help with in-house builds for AI work.

What Amazon designs for the end of a deployment

Amazon said on 30 June 2026 that its own AWS forward deployed engineering model is designed so customers are self-sufficient when a deployment ends [6]. That is a goal any company can write into a contract with any firm: the engineers leave, and your own staff can run and change the system. If you hire a firm, ask how it will meet that goal. If you hire your own engineers, the knowledge stays in your company for as long as they do.

How much time your own FDEs spend with users

If you hire your own, decide how their time is split. First Round Review reported in February 2026 that at Serval, a software company, FDEs are expected to spend "around 20%" of their time with customers and to focus on building product features [7]. Inside a company that buys software, the users are your own staff, and the split may be different. Write it into the job description.

The two paths side by side

Question	Your own FDEs	An FDE firm
Who employs the engineers	You	The firm, or its contractors
What you pay	Salary, benefits, recruiting and management	The firm's rate, which covers those costs plus its margin
How soon work can start	After a search you run yourself	After the contract, with people the firm already has
Who replaces an engineer who leaves	You	The firm, under the contract
Who keeps the knowledge	Your company	The firm, unless the contract requires a handover
How the work ends	It continues while the engineers stay	At the end date, with a handover

When hiring your own makes sense

Hire your own forward deployed engineers when the work will last for years, when the systems hold data that must stay inside your company, and when you can wait for the hiring search to finish. You also need someone senior who can judge an FDE's work in an interview. The FDE guide's page on how to hire a forward deployed engineer covers that step.

When a firm makes sense

Hire a firm when you need people soon, when the work has a clear end, when you need several skills for a short time, or when nobody on your staff can judge an FDE's work yet. Contract for a handover from the first day, so the knowledge stays when the firm leaves.

The mixed path

The two paths can be combined. A firm starts the work while you hire. As your own engineers join, the firm's engineers pair with them and then leave on a set date. The FDE guide's page on handover and exit explains what a handover should contain.

What the firms on this list say about staying and leaving

Each firm describes the length of its work in its own words.

• Reveneau says it takes responsibility for the whole project, through production and after release.
• Fluxon says its builders "stay for as long as you need them", and that "What we leave behind should feel like your own" [8].
• Factored says its more than 500 engineers let it "expand teams as your systems, priorities, and technical requirements evolve" [10].
• 10Pearls offers one senior FDE, several engineers as one unit, or a team that stays in place [9].
• Tenex says, "We become an extension of your team, building whatever you need" [14].
• Artium says it brings forward deployed engineers to build AI agents and to change how a client's own engineering teams work [13].
• HatchWorks AI says its FDEs build with Claude and "train your organization" to keep the new way of working [11].

Ask each firm the same three questions. How long does a typical engagement of this size run? What is the handover plan? And what will your own staff be able to do on the day after the firm leaves?

Questions to decide between the paths

1. How long will this work last: months, or years?
2. Can you wait for a hiring search, and who will run it?
3. Does anyone on your staff know how to judge an FDE's work?
4. Must the data and the knowledge stay inside your company?
5. What will your staff need to run on their own when outside help ends?

The full list of FDE firms shows who each firm fits, and what forward deployed engineering costs from a US firm covers the price questions. To discuss a project with Reveneau, contact Reveneau.