FDE vs Solutions Engineer vs AI Engineer: The Real Differences

Quick answer: A Solutions Engineer works mostly before the contract is signed: demos, proofs of concept, scoping, and convincing a buyer the product can do the job. A Forward Deployed Engineer works mostly after it is signed: writing production code inside the customer's environment and owning whether the thing actually works. An AI Engineer is usually not customer-facing at all: they build AI features on top of existing foundation models, in their own company's product. The clean test is what you are accountable for. The SE is accountable for the customer believing it will work, the FDE for it working in the customer's systems, and the AI Engineer for it working in the product.

Why these three titles get confused

All three roles involve engineers, AI, and a lot of talking to people who are not engineers. Job descriptions blur them further, because "AI Engineer" is currently the most searched title in tech hiring and companies attach it to work that is really data science, platform engineering or plain backend development. Meanwhile "Solutions Engineer", "Sales Engineer", "Solutions Architect", "Customer Engineer" and "Forward Deployed Engineer" get used almost interchangeably by recruiters who are describing four genuinely different jobs.

The result is that people accept offers expecting one job and get another. An engineer who wanted to build takes a Solutions Engineer role and spends the year in pre-sales calls. Someone who enjoys customers takes an AI Engineer role and never speaks to one. The titles will not protect you. Reading the responsibilities will.

This guide separates the three properly. If you only want the Forward Deployed Engineer role in depth, we have a longer Forward Deployed Engineer career guide covering the skills, roadmap and portfolio projects.

What each role actually is

Solutions Engineer

A Solutions Engineer is the technical half of a sales team. When a prospective customer asks whether the product can integrate with their existing stack, handle their data volumes, or meet their compliance requirements, the SE is the person who answers credibly and then proves it. The work is demos, proofs of concept, technical discovery, security questionnaires, architecture diagrams and RFP responses.

The output is not usually production code. It is a customer who believes, with technical justification, that this will work for them. Sales engineers, pre-sales engineers and customer engineers are broadly the same job with different titles.

Forward Deployed Engineer

A Forward Deployed Engineer embeds with a customer after the deal closes and builds the thing in their environment. The role was popularised by Palantir and has been adopted by AI labs: Anthropic's own Forward Deployed Engineer posting describes embedding "directly with the company's most strategic customers" to build production applications using Claude inside customer systems, including MCP servers and agents, with around 25% travel to customer sites (Anthropic job board).

The output is a working system inside somebody else's infrastructure, with the FDE accountable when it breaks. This is the crucial difference from the SE: the FDE is still there, on the hook, months after the signature.

AI Engineer

An AI Engineer builds AI-powered features into a product, usually on top of existing foundation models rather than training their own. The day looks like integrating an LLM API, designing prompts and agent flows, building retrieval over company data, writing evaluations, controlling token cost and latency, and shipping the result behind a feature flag.

The distinction worth holding onto: an ML Engineer typically trains, deploys and monitors models on proprietary data, while an AI Engineer typically builds on models somebody else trained. Both titles are applied loosely, so read the responsibilities rather than the heading. If you want the India-specific route into this role, see our AI Engineer roadmap for India.

The comparison, dimension by dimension

Dimension	Solutions Engineer	Forward Deployed Engineer	AI Engineer
Where in the deal	Before the contract	After the contract	Not in the deal at all
Accountable for	The customer believing it will work	It working in the customer's systems	It working in your own product
Writes production code	Rarely, mostly demo code	Yes, in the customer's environment	Yes, in your company's codebase
Customer contact	Constant, many customers at once	Constant, few customers, deep	Usually none
Who they sit with	Sales team	Applied AI, delivery or field engineering	Product or engineering team
Typical AI depth	Conceptual, enough to be credible	Applied, enough to make it work in production	Applied to deep, depending on the team
Travel	Some, for key accounts	Often significant, on-site with customers	Little to none
Ambiguity tolerance	High, requirements shift mid-cycle	Very high, the brief is often wrong	Moderate, there is usually a spec
How pay is structured	Base plus variable tied to sales targets	Usually straight engineering compensation	Straight engineering compensation
Fails when	You cannot read a room	You cannot ship without a clean spec	You cannot evaluate your own output

The boundary people get wrong: SE vs FDE

If you remember one thing, make it this. The handoff point is the signature.

The Solutions Engineer's job is finished, in the main, when the customer signs. Their proof of concept was built to demonstrate that something is possible, not to survive real traffic. That is not a criticism; it is the correct design of the role. A POC that took three weeks to harden would have lost the deal.

The Forward Deployed Engineer's job starts roughly where the SE's ends. They inherit the promise the SE made and have to make it true inside a real environment, with the customer's actual data, actual permissions, actual legacy systems and actual users who will do things nobody anticipated. When it breaks at 2am, it is the FDE's problem.

This is why the two roles attract different people. The SE optimises for persuasion and breadth: many customers, many conversations, fast context switching. The FDE optimises for delivery and depth: few customers, months at a time, owning the mess. Both need to talk to people, which is why they get confused. They are accountable for entirely different things.

There is also a consulting comparison worth making: a consultant typically delivers a recommendation, while an FDE delivers something that runs and then keeps owning it.

Where the AI Engineer genuinely differs

The AI Engineer is the odd one out, and not because of the AI. Both other roles now involve plenty of AI work. The difference is the customer.

An AI Engineer's users are usually anonymous and numerous: everybody who uses the product. They ship a feature and watch aggregate metrics. They rarely sit in a room with the person whose workflow they changed. The feedback loop is dashboards, evals and support tickets.

An FDE's users are specific and named. They can watch one person use the thing they built and see exactly where it fails. That is a much tighter feedback loop and, for some engineers, far more satisfying. For others it is exhausting, because the person is right there and unhappy.

A useful way to choose: do you want to build one thing for thousands of people, or many things for a handful of people? The AI Engineer does the former. The FDE does the latter. The SE does neither; they explain the thing to hundreds of people before it exists for any of them.

What a week actually looks like

	Solutions Engineer	Forward Deployed Engineer	AI Engineer
Most of the week	Customer calls, demos, discovery sessions	Writing code inside a customer's environment	Writing code in your own repository
Recurring pain	A prospect asks something the product cannot do	The customer's data is nothing like the sample data	The model does the right thing 92% of the time
A good week looks like	A deal moves to the next stage	Something you built goes live for real users	An eval score moves and cost per request drops
Who you argue with	Your own product team, about the roadmap	The customer's IT and security teams	Your PM, about scope and latency budgets

How the interview loops differ

This is where candidates most often prepare for the wrong job.

Solutions Engineer: expect a mock demo or a technical presentation to a fake customer. They are assessing whether you can explain something technical to a sceptical buyer, handle an objection without becoming defensive, and stay credible when you do not know an answer. Coding rounds exist but are usually lighter.
Forward Deployed Engineer: expect real coding, plus a scenario round about an ambiguous customer situation. Anthropic's posting asks for 4+ years in a technical, customer-facing role and production experience with LLMs including prompt engineering, agent development and evaluation (Anthropic job board). The scenario round is the one people fail: they propose a perfect architecture instead of asking what the customer actually needs by Friday.
AI Engineer: expect coding, system design, and increasingly a practical LLM round: build a small retrieval or agent flow, then explain how you would evaluate it and control its cost. Being able to say how you would know your output is wrong matters more than knowing model trivia.

What these roles pay, and what we could not verify

We will be precise about what is verified here, because compensation is the area where career content is least reliable.

Verified. Anthropic's Forward Deployed Engineer posting states an annual salary of $280,000 to $320,000 USD for roles based in New York, San Francisco and Seattle (Anthropic job board). That is a US frontier-lab number and should not be read as typical for the title anywhere else.

Not verified. India-specific compensation for the Forward Deployed Engineer title is genuinely thin, because the title is new here and the sample sizes behind aggregator averages are small. Published figures disagree sharply with one another, with some sources quoting entry bands around ₹8 to 15 LPA and others quoting ₹18 to 28 LPA for what they describe as the same level. We are not going to pick one of those and present it as fact. If someone shows you a confident FDE salary band for India, ask how many salaries it is based on.

Structurally different. One compensation point is not about the amount at all. Solutions Engineer pay is typically split into a base salary plus a variable component tied to sales targets, often quoted as OTE, or on-target earnings. FDE and AI Engineer roles are usually straight engineering compensation. This matters more than the headline number: a quoted OTE assumes quota attainment you do not control, so when comparing an SE offer against an engineering offer, compare base against base first.

For what Fireblaze alumni have actually been placed at, with methodology, see our placements page. We publish verified numbers only.

Which one should you choose?

Ignore prestige. These roles suit genuinely different temperaments.

Choose this	If this sounds like you
Solutions Engineer	You enjoy explaining and persuading more than building. You like variety and many conversations. You are comfortable with your success being partly tied to a sales number.
Forward Deployed Engineer	You want to build, but you find pure product work too far from the people using it. You are calm when requirements are unclear. You do not mind travel, mess, or owning something in production.
AI Engineer	You want depth in the technology itself, long uninterrupted build time, and a clean codebase you control. You would rather read an eval report than sit in a customer meeting.

A blunt filter: if the thought of a customer watching you debug their production system makes you anxious rather than energised, the FDE role is not for you, whatever it pays.

Moving between the three

These roles are more connected than the titles suggest, and moving between them is common.

SE to FDE: the most common move, and the customer skills transfer directly. The gap to close is production ownership: demo code and code that runs unattended for a year are different disciplines. Build and operate something real before you interview.
AI Engineer to FDE: the technical bar is usually already met. The gap is customer craft: discovery, saying no diplomatically, and resisting the urge to build the elegant version when the customer needs the working version.
FDE to either: the easiest direction. FDEs generally have both the technical and the customer half, so they can move into product engineering or into senior pre-sales without a step back.
Developer to AI Engineer: the shortest route if you already write production code, because it adds a layer (models, prompts, retrieval, evals) rather than replacing your existing skill set.

The India picture

A realistic note, because most of our readers are in India. Dedicated Forward Deployed Engineer roles here are concentrated in a small number of places: global capability centres of enterprise software firms, AI-heavy startups selling to enterprise, and Indian offices of US companies with large customers here. The hiring volume is genuinely smaller than for AI Engineer roles.

Solutions Engineer roles are the most numerous of the three in India, because every enterprise software vendor needs pre-sales technical staff, and it remains an underrated entry point for engineers who are good with people.

AI Engineer roles are the fastest growing of the three, though the title is applied so loosely that you must read the responsibilities to know what you are applying for. Our analysis of the India AI job market covers what the data does and does not show.

If you already write code and want to move specifically toward forward deployment work, our Applied AI Forward Deployment Engineering Program is built for working IT professionals with one to eight years of experience. If you are earlier in the journey and want the underlying AI engineering foundation first, the Artificial Intelligence course is the better starting point.

Sources and limitations

Anthropic, Forward Deployed Engineer job posting, salary range, required experience and responsibilities: job-boards.greenhouse.io/anthropic. Verified directly at the time of writing.
Role boundary descriptions are drawn from published job descriptions across AI and enterprise software companies, and from the way these teams are structured in practice.
We did not publish India salary bands for the Forward Deployed Engineer title because the available figures disagree materially and the underlying sample sizes are small. Aggregator pages for this title exist; we could not verify their methodology, so we have not repeated their numbers as fact.
Compensation, titles and role boundaries change. This piece was written on 30 September 2026.

FAQ

Frequently Asked Questions

What is the difference between a Forward Deployed Engineer and a Solutions Engineer?

Timing and accountability. A Solutions Engineer works before the contract is signed, running demos and proofs of concept to help close the deal. A Forward Deployed Engineer works after it is signed, writing production code inside the customer's environment and owning whether it keeps working.

Is an AI Engineer the same as a Forward Deployed Engineer?

No. An AI Engineer builds AI features into their own company's product and is usually not customer-facing. An FDE builds inside a specific customer's systems and spends much of the week with that customer. Both write production code, but for different audiences.

What is the difference between an AI Engineer and an ML Engineer?

An ML Engineer typically trains, deploys and monitors models on proprietary data. An AI Engineer typically builds on models somebody else trained, integrating LLM APIs, retrieval, agents and evaluations. Both titles are used loosely, so read the listed responsibilities rather than the heading.

Which role pays the most?

It depends far more on company and location than on the title. Anthropic's Forward Deployed Engineer posting states $280,000 to $320,000 USD for US roles, but that is a frontier-lab figure. Solutions Engineer pay usually includes a variable component tied to sales targets, so compare base salary against base salary first.

Do Solutions Engineers write code?

Some, but usually demo and prototype code rather than production systems. The output of the role is a customer who believes, with technical justification, that the product will work for them. Code is a means to that, not the deliverable.

Is Forward Deployed Engineer a sales role?

No. FDEs are customer-facing, which causes the confusion, but they are delivery engineers rather than sales staff. They are typically compensated as engineers, not on quota, and their work begins where the sales cycle ends.

Which role is best for someone who does not like talking to customers?

AI Engineer. It is the only one of the three where customer contact is usually minimal. Both Solutions Engineer and Forward Deployed Engineer roles involve constant direct contact with customers, and neither is enjoyable if that drains you.

Can I move from Solutions Engineer to Forward Deployed Engineer?

Yes, and it is the most common of these transitions. The customer-facing skills transfer directly. The gap to close is production ownership: build and operate something real, because demo code and code that runs unattended for a year are different disciplines.

Can an AI Engineer become a Forward Deployed Engineer?

Usually yes on technical grounds. The harder part is customer craft: running discovery, saying no diplomatically, and shipping the version the customer needs this week rather than the elegant version you would prefer to build.

Which of the three has the most job openings in India?

Solutions Engineer roles are the most numerous, because every enterprise software vendor needs pre-sales technical staff. AI Engineer roles are growing fastest. Dedicated Forward Deployed Engineer roles are the smallest group and are concentrated in global capability centres and enterprise-focused AI startups.

Do these roles require a machine learning degree?

Generally no. FDE and AI Engineer roles ask for production engineering ability plus applied AI skill, such as prompt engineering, retrieval, agents and evaluation. Research-level ML depth matters more for ML Engineer and research roles than for any of these three.

How much travel does a Forward Deployed Engineer do?

It varies by employer and account. Anthropic's posting estimates around 25% travel to customer sites. Some FDE roles are heavier, particularly when embedding with a single large enterprise customer, and some are largely remote.

What do interviewers actually test for a Forward Deployed Engineer role?

Real coding, plus a scenario round about an ambiguous customer situation. The scenario is where most candidates fail, usually by proposing an ideal architecture instead of first establishing what the customer needs delivered and by when.

Is Solutions Engineer a good entry point for engineers?

It is an underrated one, particularly if you are strong with people and enjoy explaining technology. It builds customer judgement quickly, and it opens a clean path into forward deployment or product roles later, provided you keep your hands-on coding current.

Are these three roles going to be automated by AI?

All three involve AI heavily and all three are changing because of it, but the parts that are hardest to automate are the parts that define them: understanding what a customer actually needs, deciding what is worth building, and owning the result when it breaks.

Where can I learn the skills for the Forward Deployed Engineer role?

Fireblaze AI School runs the Applied AI Forward Deployment Engineering Program for working IT professionals with one to eight years of experience, covering production Python, integration, RAG and agents, deployment and the customer-facing side of shipping AI. Our longer career guide sets out a self-study route as well.