What Is a Forward Deployed Engineer?
A forward deployed engineer is a software engineer who works inside a client's operation, on its real documents, systems and exceptions, and builds and runs a system there until it works in production. The same person learns how the work is done, builds around it and stays accountable once the system is live.
- Published
- Sources checked
- Reading time
- 8 min
- Written by
- Mirage Metrics team
Scope: what the role is, what the work looks like day to day, the skills it calls for, how it differs from four neighboring roles, and when a company needs one. It describes the role as Mirage Metrics practices it, with examples from its published deployments. It does not cover the history of the term or how other companies use the title.
Decision snapshot
- Best suited to
- Operations whose work moves between documents and several systems, with exceptions that only show up in production: order entry, freight quotes, quantity takeoff, fuel and fleet records.
- Inputs needed
- Time with the operational team, access to the systems involved (ERP, TMS, WMS, document systems), a sample of real documents, and the rules the team applies today, written down or not.
- What can be automated
- The repetitive reading, matching, checking and keying between documents and systems, once the engineer has mapped the rules and the exceptions with the team.
- What requires human approval
- Exceptions, anything outside the rules the team signed off, and the final say on what goes out: a quote, an order, a bid quantity.
- When this approach is insufficient
- A workflow nobody on the client side will own, a process with no stable rules at all, or a need a standard product already covers with configuration alone.
What a forward deployed engineer is
A forward deployed engineer, often shortened to FDE, is an engineer whose place of work is the client's operation rather than the provider's office. They write production code, but they write it against one client's documents, systems and habits, and they stay until the system holds up in daily use.
The word that matters is deployed. A model or a product that works in a demo is not yet a deployment. A deployment reads the documents the client actually receives, has access to the right systems with the right permissions, writes its result back where the team works, routes the cases it cannot settle to a person, and has someone accountable when it goes wrong. Most of that work is specific to one operation, and it is the work a forward deployed engineer does.
The same idea applied to a whole company is forward deployed engineering: a delivery model in which the provider's engineers are responsible for the system working inside the client's operation, not only for handing it over. The forward deployed engineering page sets out the model, how it differs from consulting and integration, and how a deployment is measured.
What the work looks like day to day
No two weeks look alike, but the work falls into a few recurring kinds. Most of them happen next to the people who do the work today, not at a desk away from them.
- Watching the work. Sitting with the team that handles the documents, to see how each one is actually read, checked and keyed, including the checks nobody wrote down.
- Mapping it. Writing down the workflow with that team: its exceptions, its data sources, the manual validations and the points where a person has to decide.
- Building the integration. Reading from and writing back into the client's ERP, TMS, WMS or document system, with the permissions the client grants and nothing more.
- Debugging on real inputs. An extraction pipeline that fails on scanned invoices, an order with no unit, a drawing revision that changes a quantity.
- Running the system. Watching what it flags in production and why, and adjusting the rules with the team on live data.
- Reporting. Presenting the results of a deployment to the client's decision makers, and agreeing on the next workflow.
The role travels with the work. At Mirage Metrics, engineers frequently go to operational sites, depending on projects.
The skills the role calls for
The role asks for two profiles in one person: an engineer who can ship production code, and someone the client's operational team trusts enough to show how the work really runs.
Technical depth
Strong fundamentals in software engineering, data systems or AI. In practice that means being able to debug a broken pipeline, build a custom integration with a system that was never designed for one, or improve a model that misreads a class of documents.
Operational curiosity
Wanting to understand how an operation actually runs before building anything. The valuable rules are rarely in the documentation: they are in the habits of the person who has keyed the same orders for years.
A bias for building
Preferring a working system on real inputs to a perfect architecture on paper. A forward deployed engineer is judged on what runs in production, not on the specification.
Comfort with ambiguity
Working without a clear specification, a structured roadmap or a predefined playbook, and figuring things out with the client's team as they go.
Ownership with the client
Owning the outcome rather than the output: if a deployment fails, debugging it; if a client does not understand, explaining it. There is no hand-off to another team.
How the role differs from four neighboring roles
The forward deployed engineer shares skills with several roles, which is why the title is often confused with them. What separates them is what each is accountable for and how long they stay with one client.
| Role | What they mainly do | Production code | Time with one client | Accountable for |
|---|---|---|---|---|
| Forward deployed engineer | Builds and runs a system inside one client's operation, from observing the work to production and on to the next workflow | Yes, against the client's own documents and systems | The length of the deployment, then the next workflow | The workflow running in production |
| Solutions engineer (pre-sales or sales engineer) | Shows how a product fits a prospect's need: demos, technical answers, a proof of concept during the sale | Mostly demos and prototypes, rarely the production system | The sales cycle | A credible technical case for the purchase |
| Consultant | Analyzes the operation and recommends what to change: a diagnosis, a target process, a roadmap | Usually not | The length of the engagement | The quality of the recommendation |
| Software engineer (product) | Builds features of one product for every customer at once | Yes, in the product's codebase | Little direct client time; works from aggregated feedback and the roadmap | The product's quality and roadmap |
| Support engineer | Diagnoses and resolves issues customers raise on a product already in use | Fixes, scripts and configuration; product bugs go to the engineering team | One ticket or incident at a time | Resolving the incident |
These are the usual shapes of each role. Titles vary from one company to the next, and many engineers move between them over a career. The useful question when reading a job description is the last column: who answers for the system once it is in production.
When a company needs one
A company needs a forward deployed engineer, its own or a provider's, when three things are true at the same time:
- The work moves between documents and several systems, and someone keys data from one into another by hand.
- The rules live partly in people's heads: informal checks, customer-specific conventions, exceptions handled from memory.
- Nobody on the team has the time or the profile to build an AI system and then run it alongside daily operations.
It needs one less when a standard product already covers the workflow with configuration alone, or when the open question is still what to change rather than how to run it. A product, or a consultant, is then the better fit.
A test before hiring or buying
Ask who will answer for the system six months after go-live, when a new customer format arrives or a rule changes. If the answer is nobody, the project needs a forward deployed engineer, whichever side of the contract they sit on.
How forward deployed engineers work at Mirage
Mirage Metrics is built on this model. Its engineers deploy and run AI agents inside client operations in logistics, construction, mining and manufacturing, connected to the systems already in place, more than 200 so far. A person on the client's side reviews the exceptions and signs off.
The Mirage deployment method
Five steps, in this order, for every deployment in every sector: Observe, Map, Build, Run, Expand.
Observe
Engineers spend time with the operational team before writing anything, to see how the work is actually done.
What it settles
- How the team really works
- Where friction appears
- Which decisions need a person
Map
Workflows, exceptions, data sources, manual checks and escalation points are mapped with the team.
What it settles
- The workflow to start with
- Its exceptions
- Its systems and data sources
Build
The system is built around the constraints observed on site, not around an idealized process diagram.
What it settles
- Connectors into the systems in place
- The client's own rules
- What goes to a person
Run
It runs in production with human supervision and feedback loops from day one.
What it settles
- Exceptions reviewed by the team
- Rules adjusted on live data
- Who signs off
Expand
Once a workflow is trusted, the same team moves to the next one.
What it settles
- The next workflow
- The same engineers
- The same platform
The clocks are the same for every client: an answer to every request within two business hours, a proof of concept on the client's own documents one week after the first meeting, and the first workflow in production in 4 to 6 weeks. Workflows spanning several systems can take longer and are scoped at the first meeting.
What it has produced
- John Cockerill, Morocco operations: a piping takeoff module that reads isometric drawings and produces separate MTO and BOM outputs, reviewed by the group's engineers. 80 to 90% less time spent on piping takeoff. See the case study.
- OCBM, a construction company in Casablanca: engineers worked on its real projects and drawings, starting with quantity takeoff, then bid pricing and drawing consistency. Three agents in production, every quantity traceable to its source drawing. See the case study.
- Biodéal, an organic dairy distributor: customer orders received by email, in every format, created in SAP Business One with Biodéal's own rules. A median of 48 seconds from order email to order ready, about 7% of orders flagged for a person. See the case study.
- OFRET Groupe, a freight forwarding group: two quoting agent environments, one per business line, built on site and connected to the group's TMS. More than 90% of standard quote requests answered automatically. See the case study.
- Transwin, a logistics operator in Morocco: one document workflow first, then fuel, mining equipment and a control room. Five systems live, 70% less manual data entry on freight documents, 12% lower fleet diesel consumption. See the case study.
Mirage hires for the role in Paris, Barcelona and Casablanca. The Forward Deployed Engineer role page lists the openings and how to apply.
Questions readers ask
What does a forward deployed engineer do?
They work inside a client's operation to put a system into production and keep it running: they observe how the work is done, map the workflow and its exceptions, build the integrations and rules, run the system under human supervision, then move on to the next workflow.
What does FDE stand for?
Forward deployed engineer, for the person, or forward deployed engineering, for the delivery model in which the provider's engineers work inside the client's operation and stay accountable for the system there.
Is a forward deployed engineer a software engineer?
Yes, in the sense that they write production code. The difference is where and for whom: a product software engineer builds one product for every customer, while a forward deployed engineer builds and runs a system inside one client's operation, on its own documents and systems.
What is the difference between a forward deployed engineer and a solutions engineer?
A solutions engineer usually works before the sale, showing how a product fits with demos and proofs of concept. A forward deployed engineer works after it, building the system that goes into production and staying accountable for it while it runs.
What is the difference between a forward deployed engineer and a consultant?
A consultant usually recommends what to change and leaves the implementation to the client or a third party. A forward deployed engineer builds the change, puts it into production on the client's systems and keeps running it.
What skills does a forward deployed engineer need?
Strong fundamentals in software engineering, data systems or AI; the ability to debug a broken pipeline or build a custom integration; curiosity about how operations actually run; and comfort working without a clear specification, directly with the client's team.
Does Mirage Metrics hire forward deployed engineers?
Yes. Its careers page lists Forward Deployed Engineer roles, full time, in Paris, Barcelona and Casablanca, and a six-month Forward Deployed Engineer internship in the same three cities.
Sources and scope
Checked on . Scope: the role as Mirage Metrics practices it. The comparison with other roles describes their usual shape, not any particular employer. Every Mirage figure comes from the published case study listed here.
- Mirage Metrics, John Cockerill case study: the piping takeoff module and the 80 to 90% reduction in piping takeoff time
- Mirage Metrics, OCBM case study: three agents in production and every quantity traceable to its source drawing
- Mirage Metrics, Biodéal case study: the 48-second median from order email to order ready in SAP Business One and the share of orders flagged for a person
- Mirage Metrics, OFRET Groupe case study: more than 90% of standard quote requests handled automatically
- Mirage Metrics, Transwin case study: five systems live, 70% less manual data entry on freight documents and 12% lower fleet diesel consumption
- Mirage Metrics, careers page: the Forward Deployed Engineer openings, the typical week and what the company looks for
Related on this site
- Forward deployed engineering: the model, the method, and how a deployment is measured
- Forward Deployed Engineer at Mirage: the openings in Paris, Barcelona and Casablanca, and how to apply
- Build, buy or forward deploy: three ways to get AI into an operation, compared
- The Mirage platform: the operating layer the engineers deploy
- Case studies: every published deployment and its figures
- About Mirage Metrics: the company, its founders and its offices
Get an automation plan from the engineers who would build it
Mirage Metrics deploys AI agents inside operations in logistics, construction, mining and manufacturing, with its own forward deployed engineers. Tell us which workflow costs your team the most time, and we send back a written plan.
Prefer to talk it through? Book a call