Is a forward-deployed engineer an AI engineer?

Increasingly the same person, though the roles start from different places. 6.9% of FDE postings in Plank's census also use AI-engineer or applied-AI language in the title. The durable difference is environment: an AI engineer works in a codebase they control, and a forward-deployed engineer works in one they do not.

Do forward-deployed engineers need machine learning expertise?

Deep model training expertise, rarely. Working fluency with the applied stack — retrieval, agents, evaluation, context design, inference cost — is now close to mandatory, because that is what is being deployed.

Which role is harder to hire for?

Forward-deployed, by most employers' account. The AI engineering skill set is a subset of what the job needs; the scarce part is combining it with the willingness to work inside someone else's constraints and stay accountable in front of a customer.

Forward-deployed engineer / comparisons

Forward-deployed engineer vs AI engineer: what's the difference?

An AI engineer builds AI capability — agents, retrieval, evaluations, model plumbing — usually inside their own company's product. A forward-deployed engineer takes capability into someone else's environment and makes it survive contact with their data, their systems, and their users. The distinction is not the technology, which is increasingly the same; it is whose codebase and whose constraints you are working in.

Side by side

	Forward-deployed engineer	AI engineer
Whose codebase	The customer's, plus the vendor's	Their own company's
Whose constraints	The customer's data, security posture, and legacy systems	Their own stack and standards
Success measured by	One customer reaching a working deployment	A capability shipping to all users
Integration surface	Systems of record nobody documented	Internal services and a known schema
Customer contact	Continuous and direct	Usually mediated by product management
Hardest part	Everything around the model	The model behaviour itself

What a AI engineer actually does

An AI engineer builds software on top of models: prompting and context design, retrieval pipelines, agent orchestration, evaluation harnesses, inference cost and latency work. The role is a product engineering specialty, and the customer is usually the company's own product.

Where they overlap

This is the fastest-closing gap of any pair on this list. The AI stack is now the same on both sides — agents, retrieval, evals, tracing — and the AI-native forward-deployed engineer is simply an AI engineer who works in the customer's environment rather than their own. What still separates them is tolerance for mess: an AI engineer usually controls the schema, the deploy pipeline, and the definition of done, and a forward-deployed engineer controls none of the three.

What the hiring data shows

The two are visibly merging in the postings. 68 FDE roles in Plank's census — 6.9% — also carry applied-AI, AI-engineer, ML-engineer, or agent-builder language in the title. That is the highest rate at which the FDE title fuses with any adjacent role in the dataset — twice the rate of the next one (data and platform engineering, 3.5%) and four times that of solutions engineering (1.6%). Employers including Brex, Smartsheet, ICEYE, Reflection AI are hiring for one job under both names.

From Plank’s first-party census of 982 verified open FDE roles across 462 companies, observed 2026-09-04. The census covers FDE-titled postings, so it describes the forward-deployed side of this comparison; it is not a survey of ai engineer roles.

Which one do you need?

Forward-deployed engineer

The AI works in your demo and fails in the customer's environment.
Every deployment needs bespoke integration with systems you do not control.
Your enterprise customers need someone on their side of the wall.

AI engineer

You are building AI capability into your own product.
The hard part is model behaviour, evaluation, and cost.
There is no external customer environment in the loop.

Questions people ask

Is a forward-deployed engineer an AI engineer?: Increasingly the same person, though the roles start from different places. 6.9% of FDE postings in Plank's census also use AI-engineer or applied-AI language in the title. The durable difference is environment: an AI engineer works in a codebase they control, and a forward-deployed engineer works in one they do not.
Do forward-deployed engineers need machine learning expertise?: Deep model training expertise, rarely. Working fluency with the applied stack — retrieval, agents, evaluation, context design, inference cost — is now close to mandatory, because that is what is being deployed.
Which role is harder to hire for?: Forward-deployed, by most employers' account. The AI engineering skill set is a subset of what the job needs; the scarce part is combining it with the willingness to work inside someone else's constraints and stay accountable in front of a customer.

Still working out the definition itself? Start with what a forward-deployed engineer is, or see who is hiring them and for how much.