Forward deployed engineering, explained
What a forward deployed engineer does, why the role matters for AI implementation, and how it differs from consulting, outsourced development and on-site support.
AI projects fail in a predictable place: between the strategy document and the codebase. The people who understand the business are not the people writing the code, and every handoff loses detail. By the time a requirement reaches an engineer, the exception that mattered most has often disappeared.
Where the gap comes from
- Business problem
- Workshop
- Requirements document
- Specification
- Build
- Launch
In a traditional delivery model, a strategy team runs workshops, writes requirements and hands them over. An engineering team builds to the specification. When the system meets real users, it turns out the specification missed how the work actually happens. Everyone did their job. The outcome still fails.
What the role does
- Studies the real workflow
Maps the current process by following real cases with the people who do the work, not by reading the procedure.
- Finds the data and the exceptions
Identifies which systems hold what, how clean the data is, and which edge cases the design must survive.
- Prototypes with actual users
Puts a working version in front of the people who will use it within days, and changes it based on what they do, not what they say.
- Designs checkpoints and integrations
Decides where a person must approve, how the system hands off, and what it needs to connect to.
- Challenges the request
Pushes back when a request would automate a bad process, and proposes the simpler fix when there is one.
- Stays through deployment
Remains accountable for the system working in production, not just for the code being delivered.
How it differs from other models
| Model | What you get | Who owns the outcome |
|---|---|---|
| Strategy consulting | A recommendation and a roadmap | You, once the consultants leave |
| Outsourced development | Software built to your specification | You, if the specification was right |
| On-site support | Help operating existing systems | Shared, but rarely the design |
| Forward deployed engineering | A deployed system shaped by the real work | The engineer and team, through launch |
The distinguishing feature is accountability across the gap. A forward deployed engineer is judged on whether the system works for the people using it, which forces them to understand the work deeply enough to design for it.
What it looks like day to day
Early in an engagement, most of the time is spent observing and mapping. Within the first couple of weeks there is usually a rough prototype in front of users. The middle of the engagement alternates between building, testing with users and adjusting the design. Towards the end, the focus moves to integration, permissions, monitoring and handover, so the system can be run and improved without the engineer in the room.
When you need one
- The problem is clear, but the right solution is not
- The workflow spans several teams or systems
- The system will depend on messy, real-world data
- Previous projects delivered software that people did not adopt
How Merge uses the role
At Merge, forward deployed engineers work alongside product managers, designers, AI specialists and QA as one team, so what is learned on the ground reaches the build the same week. The engineer who mapped the workflow is part of the team that ships the system, which is what keeps the build tied to the outcome.
Frequently asked questions
What is a forward deployed engineer?
An engineer who works inside the customer's operating reality, discovering the real workflow, prototyping with users, connecting data and tools, and keeping the build tied to the original business outcome.
How is a forward deployed engineer different from a consultant?
A consultant typically leaves a recommendation. A forward deployed engineer builds and deploys the system, and stays accountable for whether it works for the people using it.
Why does forward deployed engineering matter for AI projects?
AI systems depend on real data, real edge cases and real user behaviour. Being close to the work lets the engineer find those early and design the system around them, instead of discovering them after launch.
Does forward deployed mean working on site?
Not always. It means working inside the customer's real workflow with their people and data. That can be on site, remote or a mix, depending on the work.
Get the next note by email
One email a month at most. Practical, from real work. Unsubscribe in one click.