# What is a forward deployed engineer?

A forward deployed engineer is a software engineer who works inside a client's business, alongside the people doing the work, to understand a problem first-hand, build a solution and get it into production. Unlike a consultant, they write the code. Unlike a contractor, they decide what to build. The role has become common in AI projects, where the useful solution is usually found in a team's daily operations rather than in a requirements document.

By Muhammad Ali Khoja, CSO. Published 2026-10-14.

## Where did the term come from?

The title was popularised by data companies that discovered their software only delivered value once an engineer sat with the customer and shaped it around real workflows. The engineer was "forward deployed" in the military sense: stationed where the action is, with authority to act, rather than waiting at headquarters for instructions.

The idea has since spread across enterprise software and especially into AI, for a simple reason: the gap between a demo and a working system is almost always a gap in understanding how the business really works.

## How is it different from a consultant or a contractor?

|  | Consultant | Contractor | Forward deployed engineer |
| --- | --- | --- | --- |
| Decides what to build | Yes | No | Yes |
| Builds it | No | Yes | Yes |
| Sits with the business | Sometimes | Rarely | Always |
| Owns the outcome | Partly | No | Yes |

A consultant hands over a plan. A contractor executes someone else's plan. A forward deployed engineer does both, in the same week, with the people who will use the result.

## Why does the model suit AI projects?

Because AI projects fail at the edges, not in the model. A pilot built from a spec runs on an export of clean data. Production runs on three systems that disagree about what a customer is, access rules written for people rather than software, and a team that has quietly developed workarounds nobody documented.

An engineer sitting inside the team sees all of that in the first week. They can pick the use case that will actually pay off, build against the real data, and notice when the "simple" automation needs a human review step because of something that happens every month-end.

## What does a forward deployed engineer actually do day to day?

- Shadows the team and maps how work really flows, including the exceptions
- Picks the one or two tasks where automation or AI will pay off fastest
- Builds a working version against live systems, not a sandbox
- Puts it in front of the people who do the work, and iterates
- Ships it with the permissions, logging and fallbacks production needs
- Stays long enough to see it run under real conditions

## When should you use one?

When the problem is clear but the solution isn't. When the work lives in operations rather than in a product roadmap. When previous projects produced a pilot that never shipped. And almost always for a first AI project, where the most expensive mistake is building the wrong thing well.

When you already know exactly what to build, project delivery is cheaper. When you need capacity rather than judgement, resource augmentation fits better.

## Read next

- [Working with an engineering partner: a guide for technology leaders](https://fetchsky.com/insights/working-with-an-engineering-partner)
- [Forward deployed engineers](https://fetchsky.com/services/forward-deployed-engineers)

Source: https://fetchsky.com/insights/what-is-a-forward-deployed-engineer