What is a forward deployed engineer?

Forward deployed engineers build software where it's used – with the customer. What the role involves, where it comes from and why it matters so much for AI projects.

A forward deployed engineer (FDE) is a software engineer who doesn't sit at headquarters but with the customer. They understand the problem on the ground, build the solution themselves and stay until it works in daily use.

Sounds simple. But it's the opposite of how many software projects run.

Where the term comes from

The role was made famous by the software company Palantir. They realised early on that their platform was powerful, but it only delivered value for a customer if someone adapted it to that customer's real processes. So they sent engineers directly to customers – "forward deployed".

Today, AI providers such as OpenAI and Anthropic also employ forward deployed engineers. The reason is the same: a strong AI model alone doesn't solve a business problem. Someone has to understand where it should be used, connect it and make sure people actually use it.

What an FDE does differently

An FDE combines three roles that are usually split across different people:

  • Consultant: they understand the process and the actual problem – not just the requirement written on paper.
  • Engineer: they build the solution themselves instead of handing it to another team.
  • Companion: they stay until the solution has landed in daily work, and adjust it when something gets stuck.
Classic consulting Agency Forward deployed engineering
Understands the process yes partly yes, on the ground
Builds it no yes yes
Stays until it runs rarely rarely yes

Why this matters so much for AI projects

Many AI projects don't fail because of the technology. They fail because nobody knows exactly which process is supposed to improve – or because the solution is built past the people who are meant to use it.

Typical patterns:

  1. A tool is bought before it's clear which problem it should solve.
  2. A prototype works in the demo, but not with real data.
  3. The solution runs, but nobody on the team feels responsible for it or understands it.

An FDE tackles exactly these points: first understand the process, then build the smallest thing that takes real work off people's plates, and test it with real cases.

Does every business need its own FDE team?

No. Large software companies have their own FDE teams because they roll out their products at many customers. Most businesses don't need that permanently – but they benefit from the same way of working when they introduce AI or automation.

That's exactly how I work: as a forward deployed engineer for businesses that don't want to build their own team for it.

In short

  • A forward deployed engineer builds software where it's used.
  • The role combines process understanding, engineering and hands-on support.
  • With AI in particular, this way of working often decides whether a project lands in daily use or gets stuck in the demo.