What does a forward deployed engineer actually do?
Job ads for the role read like a startup CTO's. Here's the week as it happens: your data, your operators, the code, and the parts of the product that get built back at head office because of what the engineer saw.
Answered in short
4 things that decide this
- 01A forward deployed engineer spends most of the week inside a customer's business. They read the data as it really is, sit with the people who run the process, and write production code on the customer's own systems. The output is a working deployment of the vendor's product, not a slide deck or a proof of concept.
- 02The job has two halves that most engineering roles keep apart. One half is integration: wiring the product into the customer's systems of record, permissions and edge cases. The other half is feedback: carrying what broke in the field back to the core product team, which is how Palantir built much of its platform.
- 03Palantir describes the role as close to a startup CTO's: small teams and end-to-end ownership of high-stakes projects. That's the appeal for engineers, and the reason some avoid it. The same public sources note heavy travel and pressure to fix customer problems fast.
- 04For a buyer, the practical meaning is simple. An FDE is the difference between a vendor that ships a model and a vendor that ships a system your staff use on Monday. Smaller companies get the same motion through an embedded senior engineer rather than a lab's FDE team.
What the role looks like day to day
It starts with your data. Palantir's own wording for its Deltas is blunt. Gergely Orosz quoted it in The Pragmatic Engineer on 12 August 2025: their 'mandate is to achieve technical outcomes for our customers.' In our own embedded engagements that begins with exports, database access and a list of every system the process touches. Most of that list was never written down.
Then comes building. Colin Jarvis set up OpenAI's forward deployed engineering team at the start of 2025. He told Orosz that 'FDEs write code directly on customer infrastructure and use customer tooling.' That sentence separates the role from a solutions architect. The code goes into your repo, under your review, and it has to get past your security team.
What's left belongs to the product. Whatever the engineer learned gets written up for the core team: which connector was missing, which permission model broke, which workflow nobody expected. Orosz reports that until around 2016 Palantir had more FDEs than 'normal' software engineers. Many moved into core product work when Palantir Foundry launched. Its field engineers were its research arm.
What the role is not
Sales engineering ends where this job starts. A sales engineer's work is done when you sign; an FDE's begins then.
Consulting is a different product. Orosz: consultants 'make one-off recommendations, whereas FDEs generally work with customers, long-term.' An FDE is judged on whether your system runs, not on whether the advice was sound.
A permanent hire it isn't. Wikipedia's definition, read 2026-09-01, has the engineer working alongside the client's staff 'for a defined period of time.' The job ends in a handover or a support deal. The good ones plan that from week one.
- AccessData, systems, and the people who run them.
- MapWhere the product meets the process.
- ShipCode in the customer's repository.
- HardenPermissions, failure paths, monitoring.
- Report backWhat the product is missing.
- Hand overRunbook and training, or stay on.
Report back is the step a contractor skips. It's why the role sits inside a product company at all.
If you're reading this because you want the job
Palantir, OpenAI, Anthropic and Amazon Web Services all hire for it. The titles vary: forward deployed engineer, forward deployed software engineer, FDSE (Wikipedia, read 2026-09-01). Public accounts agree on the trade. You get unusual ownership and a fast education in how businesses really run. You give up a quiet life. A March 2026 Wall Street Journal piece put it in its headline: 'The Hottest Job in Tech Isn't Very Glamorous.'
What transfers are the plain skills: reading someone else's data model, writing code that passes their review, and telling a customer no with a reason. This page isn't a jobs board. Each firm above lists openings on its own careers page, and Palantir's interview process is written up widely.
Related questions
01Is a forward deployed engineer a real software engineer or a support role?+
A software engineer. Palantir's own framing is that the job resembles a startup CTO's, with small teams owning end-to-end delivery. OpenAI's FDEs write production code on customer infrastructure. Support roles answer tickets about a product that exists; FDEs build the part of it that doesn't yet.
02How is an FDE different from a customer success engineer?+
A customer success engineer keeps a live deployment healthy and the customer renewing. A forward deployed engineer creates the deployment, often writing new code and new integrations to do it. They hand off to each other: the FDE leaves a running system, and success keeps it running.
03Can a company that isn't a lab or a Fortune 500 get one?+
Yes, through a partner that embeds a senior engineer rather than through a vendor's own FDE team, which is kept for its largest accounts. The method is the same: an engineer in your repository and standups, shipping into live operations, staying until someone on your side owns it. The linked page covers how that runs at small and mid-market scale.
Related
- What is a forward deployed engineer? →The definition, the Palantir origin and the three names for one role.
- FDE vs solutions engineer vs consultant →A decision table: who writes the code and how long they stay.
- Can a small or mid-size business get a forward deployed engineer? →Yes, through an embedded senior engineer. What the engagement looks like.
- Forward deployed AI engineer: what changes when the product is a model →Evals, guardrails, data plumbing and handoffs.
- Hire forward deployed engineers →A senior engineer embedded in your operations, from a free scoping call.
- Staff augmentation →Senior engineers who join your team and stay after launch.
- What does a production AI pod look like? →The team shape around an embedded engineer.
