Guide
What is a forward deployed engineer?
A forward deployed engineer (FDE) is a software engineer who works inside a customer's company to make a product do real work there. They understand the problem with the people who have it, build on the customer's data and systems, and stay until the result is in daily use.
The role
Who is a forward deployed engineer?
A forward deployed engineer (FDE) works inside a customer's company and makes an AI product work there. The job joins three kinds of work.
fig. / the forward deployed engineer sits where three jobs meet
Pick a circle to see what the job takes from it.
Writes and tests the code.
A forward deployed engineer builds real software every week, like any software engineer.
What does a forward deployed engineer do?
Four things, in order. Understand: agree in writing what done means. Build: on the customer's data, inside their systems, with the access they give. Deploy: pass their security checks and put it into real use. Stay: until people use it, then tell the product team what you learned.
from
The customer
their data / their systems / their rules
- 01
Understand
Sit with the people who have the problem. Agree in writing what done means.
- 02
Build
On their data and inside their systems, with the access they give you.
- 03
Deploy
Pass their security checks and put it into real use.
- 04
Stay
Stay until people use it. Then tell your product team what you learned.
back to
Your product team
learns what customers really need
How is a forward deployed engineer different from an AI or ML engineer?
An AI or ML engineer builds and improves models and the apps around them, usually inside their own company. A forward deployed engineer takes such a system into a customer's company and makes it work there: their data, their systems, their rules. The first owns the model. The second owns the result for the customer.
| Compare | AI or ML engineer | Forward deployed engineer | Solutions engineer |
|---|---|---|---|
| Works mostly | Inside their own company | Inside the customer's company | Between sales and the customer |
| Responsible for | The model and the app around it | The result for one customer | The fit between product and need |
| Writes code for real use | Yes | Yes | Sometimes |
| Data | Chosen and prepared by the team | Whatever the customer has | Sample or demo data |
| Success is | A better model or app | A system in daily use | A working pilot |
| Hardest part | Quality at large scale | Limits you did not set | Understanding the need |
- Works mostly
- Inside their own company
- Responsible for
- The model and the app around it
- Writes code for real use
- Yes
- Data
- Chosen and prepared by the team
- Success is
- A better model or app
- Hardest part
- Quality at large scale
What skills does a forward deployed engineer need?
Strong software engineering first: you write and release code under your own name. Then five things most engineers are never taught: turning an unclear request into a plan, working inside systems you cannot change, handling data you did not choose, passing security checks, and explaining a technical decision to people who do not code.
- Engineering
- Writing, testing and reviewing code. Everything else stands on this.
- Customer problems
- Turning what someone says they want into what they need, in writing.
- Deployment
- Passing someone else's security checks and release steps.
- Product knowledge
- Knowing what the product can do today, and what it should not be forced to do.
- AI systems
- Search, testing and hosting for AI apps, and the ways each one fails.
- Data
- Messy, incomplete and access-controlled. You work with what exists.
- Integration
- Connecting to systems you did not build and cannot rewrite.
- Fast first versions
- Something that runs by Friday, so the talk is about a real thing.
- Clear communication
- An honest status report, for people who do not code.
Career pathway
Five ways into the role.
Pick where you are now. You see what you already have, what you add and which program covers it.
- you are here
AI or ML engineer
You have: training and testing models.
- next
- then
Forward deployed engineer
You add: customer needs, customer data and setup inside the customer's systems.
View the program
Already build full AI apps? Go straight to the last step.
Is forward deployed engineering right for you?
It suits engineers who like being close to the problem. You will enjoy it if you prefer a real limit to a perfect design, if you can explain your work to people outside engineering, and if you would rather finish something useful this week than something perfect next quarter.
You will like it if
- You like variety: new systems, a new problem, a new team.
- You would rather talk to the person with the problem than read a ticket.
- You can work within limits you did not choose.
It may not suit you if
- You want to go deep on one system for years.
- You dislike switching between tasks.
- You want every requirement fixed before you start.
The program
Train for the role.
About the role
Forward Deployed Engineer / P-04
- You become
- Forward deployed engineer
- Duration
- 1 to 6 months, depending on the program.
- Format
- Online and classroom
- You need first
- Experience of building or training an AI system
- Start dates
- New batches through the year
- Job support
- 100% job enablement program
Three short steps.
View the program