A forward deployed engineer (FDE) is a software engineer who embeds with a company's customers and writes production code inside their systems until the product delivers a measurable result. Think of it as part engineer, part consultant, part product manager, with the engineering doing most of the work.
The role started at Palantir and is now one of the most in-demand titles in AI. As of November 2026, we track 44 open forward deployed engineer roles at 20 companies on our forward deployed engineer job board, with a median posted base midpoint around $243K.
What is forward deployed engineering?
Forward deployed engineering is the practice of putting engineers in the field, next to the customer, instead of behind a product roadmap. The FDE owns a customer outcome, not a feature.
Palantir, which created the role, puts the distinction neatly in its post Dev versus Delta: a product engineer's focus is "one capability, many customers," while a forward deployed engineer's focus is "one customer, many capabilities."
In practice an FDE:
- Scopes the customer's real problem, which is rarely the one in the sales deck
- Builds the integration, pipeline, agent, or app that solves it, in the customer's environment
- Ships it to production and stays until people actually use it
- Feeds what they learned back to the product and research teams
That last point separates an FDE from a services contractor. Good FDE teams turn field work into product.
Where the forward deployed engineer role came from
Palantir created the "Forward Deployed Software Engineer" title in the early 2010s and called the role "Delta" internally. Palantir says the name is a throwback to when each Business Development team was named after a letter in the NATO alphabet. Its less-technical counterpart, the Deployment Strategist, is called "Echo".
According to The Pragmatic Engineer, Palantir had more FDEs than regular software engineers until around 2016, when it launched Foundry and many FDEs moved back into product work. Palantir's customers were governments and large, traditional enterprises. Software that sensitive and that complex did not install itself, so Palantir sent engineers in to make it work.
Why AI companies hire FDEs now
Enterprise AI has the same problem Palantir had, only bigger. A model is useless to a bank until someone connects it to the bank's data, permissions, and workflows, then proves it works with evals.
Andreessen Horowitz made this argument in Trading Margin for Moat, calling the FDE "the hottest job in startups." Their thesis: AI app companies should accept lower early gross margins and embed engineers deeply, because the company that becomes the customer's system of work wins the account for years.
The AI labs agree. OpenAI's FDE posting says its FDEs "lead complex end-to-end deployments of frontier models in production alongside our most strategic customers," with travel up to 50%. Anthropic's FDE posting asks FDEs to deliver "MCP servers, sub-agents, and agent skills" used in production workflows, with 25-50% travel to customer sites.
AI apps are hiring too. On our board, Cursor (6 open roles) and Notion (5) are the biggest FDE hirers after OpenAI. Selling an AI product to a 10,000-person enterprise means someone has to roll it out, wire it into existing tools, and get adoption past the pilot. That someone is increasingly an FDE.
What a forward deployed engineer does in a typical week
No two FDE weeks look the same, but the shape is consistent. Here's an illustrative week for an FDE deploying an AI agent at a financial services customer, built from how OpenAI describes its FDE engagements (scoping, validation, delivery) and how Palantir FDEs describe theirs:
| Day | What the FDE is doing |
|---|---|
| Monday | On-site with the customer's ops team, mapping the claims-review workflow step by step |
| Tuesday | Pulling sample data, building a first prototype, writing eval cases with the team that will use it |
| Wednesday | Debugging permissions and an SSO integration with the customer's IT team |
| Thursday | Hill-climbing on evals, shipping a new version, demoing it to the workflow owner |
| Friday | Writing up field notes for product and research, turning reusable code into a template |
A Palantir FDE quoted in the same Dev versus Delta post describes it simply: "Most weeks, I spend a couple of days working at the customer premises... and the rest of the time monitoring, debugging, deploying, or configuring our software for that customer."
Forward deployed engineer skills
FDE job posts are remarkably consistent. Pulling from the OpenAI and Anthropic postings linked above:
- Production engineering. OpenAI asks for production-grade code "across frontend and backend using Python, JavaScript, or comparable stacks." Anthropic asks for strong Python, ideally plus TypeScript or Java.
- LLM deployment experience. Prompting, agents, evals, and the judgment to know when a model is the problem versus the data.
- Customer discovery. Running a room of stakeholders, finding the real problem, and saying no to scope that won't land.
- Problem decomposition. Palantir FDEs call it "technical decomp": taking a vague business problem and breaking it down to the lines of code that solve it.
- Autonomy. Small teams, unfamiliar systems, no one to unblock you.
- Domain depth (a plus). Anthropic lists financial services and healthcare backgrounds as a plus. "Financial Services" is one of the most common tags on our FDE listings.
The full skills-and-portfolio path is in our guide on how to become a forward deployed engineer.
Forward deployed engineer vs AI engineer
This is one of the most common questions about the role. The short version:
| Forward deployed engineer | AI engineer | |
|---|---|---|
| Builds for | One customer at a time | Many users of their own company's product |
| Works in | The customer's systems and data | Their own company's codebase |
| Measured on | Customer outcome and adoption | Product quality, reliability, usage |
| Customer contact | Daily, often on-site | Occasional |
| Travel | Common (25-50% in the AI lab postings above) | Rare |
An AI engineer builds AI features into a product. An FDE takes that product and makes it work for one specific customer. The skills overlap heavily, which is why plenty of AI engineers move into FDE roles and back.
FDEs also get confused with solutions engineers. The difference matters for pay and career path, so we wrote a full breakdown: forward deployed engineer vs solutions engineer.
Who is hiring forward deployed engineers
As of November 2026, our board lists 44 forward deployed engineer roles at 20 companies. The top hirers:
| Company | Open FDE roles |
|---|---|
| OpenAI | 9 |
| Cursor | 6 |
| Notion | 5 |
| Anthropic | 4 |
| Gradial | 2 |
| Lovable | 2 |
| Ramp | 2 |
| Snorkel AI | 2 |
| Replit | 1 |
| AssemblyAI | 1 |
The most common tags are Enterprise (22 roles), AI Agents, Services, Frontier Models, and Financial Services. Work style: 26 on-site, 13 hybrid, 5 remote. This is a field job, and only about 1 in 9 roles is fully remote.
Seniority skews up. Of the 44 roles, 4 are senior, 4 staff, 2 lead, 5 manager, and 1 head of function. Companies are building FDE teams from scratch and need people who can lead them. Browse them all on the FDE job board, or see every Brand Engineering role on the full board.
How much do forward deployed engineers make?
Of the 44 roles on our board, 25 list a USD salary. The median posted base midpoint is about $243K, the 75th percentile is about $260K, and the full range runs from $130K to $432K. That's the highest median of any role family we track.
Top of the range:
- Regional Director, Forward Deployed Engineering - Financial Services @ Cursor: $288K-$432K
- Manager, Forward Deployed Engineering @ Anthropic: $320K-$405K
- Forward Deployed Engineering Lead @ Lovable: $241K-$390.5K
These are posted base salary ranges, not total comp. Equity at AI labs can add a lot on top. The full breakdown, including third-party total comp data, is in our forward deployed engineer salary guide.
FAQ
What is forward deployment engineering?
Forward deployment engineering means embedding software engineers with customers so they can build and ship production solutions inside the customer's environment. The engineer owns the customer's outcome end to end, from scoping to adoption. Palantir pioneered the model, and AI labs like OpenAI and Anthropic now run their own FDE teams.
How much do forward deployed engineers make?
On our board, the median posted base midpoint for FDE roles is about $243K as of November 2026, across 25 salaried postings. Ranges go from $130K up to $432K for leadership roles. Total comp is higher at companies that grant equity. See the FDE salary guide for detail.
What is the difference between an AI engineer and a forward deployed engineer?
An AI engineer builds AI features into their own company's product for many users. A forward deployed engineer deploys that product for one customer at a time, inside the customer's systems, and is measured on that customer's outcome. FDEs spend far more time with customers and usually travel.
Is a forward deployed engineer a consultant?
No, though people mix them up. Palantir FDEs point out that consultants usually produce a one-time analysis or recommendation, while FDEs deploy their company's own software and build long-term solutions. FDEs also contribute code back to the core product.
Do forward deployed engineers need to code?
Yes. Every major FDE posting we've reviewed asks for production engineering skills, usually Python plus a frontend or second backend language. The coding is the job. Customer skills are what make it an FDE job instead of a regular software role.
Is forward deployed engineer an entry-level job?
Usually not at AI labs. OpenAI asks for 5+ years of engineering or deployment experience and Anthropic's London posting asks for 4+. Palantir is the exception: per The Pragmatic Engineer, it hires FDEs with as little as a year of post-college experience. If you're starting out, see how to become a forward deployed engineer.
