A Forward Deployed Engineer (FDE) is a developer who works directly with a customer's team, on its code and data, to get a product into production. The title comes from software vendors and AI labs, but the same work can be done part time for a small team. That's what we offer at Klyx, and we spell out below what it covers.
What is a Forward Deployed Engineer?
A Forward Deployed Engineer is, in The Pragmatic Engineer's definition, "a software engineer who alternates between being embedded with customer teams and core product engineering teams" (The Pragmatic Engineer, 12 August 2025).
Marty Cagan of SVPG stresses the purpose of being there: to understand the customer's environment, their problems, and what it really takes to solve them (SVPG, 17 September 2025). He also notes the role is widening to other profiles such as product managers and designers.
Where does the Forward Deployed Engineer role come from?
The Forward Deployed Engineer role started at Palantir in the early 2010s, where it was called "Delta" (The Pragmatic Engineer, 12 August 2025). AI labs have since picked it up to deploy their models at customers.
OpenAI, for example, is hiring in Paris. Its posting describes FDEs who lead end-to-end deployments of models in production at customers, with 3 days a week in the office and up to 50% travel (OpenAI posting on the a16z job board, posted 5 October 2026).
What does a Forward Deployed Engineer do day to day?
Day to day, an FDE writes production code inside the customer's environment. They sit in team meetings, learn the data and constraints on site, then ship. The difference from a regular developer job is how close they are to the user: the FDE sees the problem first-hand and doesn't wait for it to arrive as a ticket.
At Klyx, that takes a specific shape (FDE page, November 2026):
- We work in your Slack, your repo and your team meetings.
- We take on part of the backlog, from design to production.
- Each week you get a written update listing the PRs shipped.
- At the end of the engagement, handover documentation lets your team take over.
What's the difference between an FDE, a consultant and a freelancer?
An FDE differs from a consultant by writing and shipping production code, and from a typical freelancer by working on your current priorities and not a fixed spec. Cagan describes the opposite model at large services firms: the client provides the spec, a price gets agreed, and the provider builds what's requested, unable to commit to an outcome because the solution is dictated to them (SVPG, 17 September 2025).
The lines stay blurry, and the title guarantees nothing. When a provider calls itself an FDE, ask to see where the code lands and who reviews it.
Why is everyone talking about Forward Deployed Engineers?
People talk about the role because job postings have jumped: Bloomberry counts a 1,165% rise in postings titled "forward deployed engineer" from January to October 2025, against the same period in 2024 (Bloomberry, updated 25 January 2026). The analysis covers about 1,000 postings supplied by Revealera. A percentage that high implies a small starting base, so it's more striking than the volume.
Our reading: an AI prototype is quick to build, but connecting it to real data and tools remains engineering work done close to the customer. That's our opinion, with no figure behind it.
How much does a Forward Deployed Engineer earn?
The median FDE salary is $173,816 across the postings analysed by Bloomberry that show a range (Bloomberry, updated 25 January 2026). That figure comes from dollar-denominated postings and doesn't carry over to France. We found no reliable French salary data for this job, and we'd rather say so than quote one from an aggregator.
Can an SME hire a time-shared FDE?
Yes, as long as it has a product or a codebase to work on and someone to set priorities. Without hiring full time, a small company can bring in an FDE a few days a week. The format doesn't suit every SME: with no tech team and no repo there's nobody to join, and scoping followed by a fixed-price project fits better.
Our version is simple: Maxime Mansiet, our lead engineer, joins your team (about). The terms, from our pricing (November 2026):
- €700 excl. VAT a day.
- 2 to 3 days a week.
- 4 weeks minimum, so at least €5,600 excl. VAT (2 days a week for 4 weeks).
- Two clients at most at any time.
There's also what we don't offer: no 24/7 on-call and no SOC 2 certification. If your project needs them, we tell you on the first call.
FDE or fixed-price project: which should you choose?
Choose an FDE when the backlog is open and changes every week, and a fixed-price project when the expected result fits in one sentence. An AI Sprint, from €4,500 excl. VAT in 2 weeks, delivers one workflow with its scope set in advance (Klyx pricing, November 2026). With an FDE, you pay for time spent in your team. If you don't yet know which process to tackle, an AI audit, from €1,500 excl. VAT, answers that first (same source).
The discovery call lasts 30 minutes: you tell us about your product and what needs to ship, and we tell you which of these formats fits.