Forward deployed engineer vs solutions engineer vs sales engineer vs solutions architect

If you are trying to work out which of these people to hire, or which one a vendor is trying to sell you, the fastest way to tell them apart is to ask two questions. Does this person work before the sale or after it? And who is accountable when it doesn't work in your environment?

I have been a solutions architect for most of twenty years, I have written production code for longer than that, and for the last few years I have done forward deployed work: building AI systems inside other people's companies. I have sat through a lot of sales engineer demos from the buying side. So this is not a recruiter's summary. It is what the roles look like when you are paying for them.

The four roles side by side

The column that matters is the accountability row. Three of the four roles are gone before the system fails.

What does a solutions engineer do, and when does it stop?

A solutions engineer is the technical person a software vendor brings to the sales process. They understand the product deeply, they can build a proof of concept against a sample of your data, and they can answer the questions your IT team asks that the salesperson can't. Good ones are genuinely useful. They will tell you honestly whether the product fits.

The limit is timing. Their job ends when you sign. The proof of concept was built on clean sample data, in the vendor's environment, without your access rules, your legacy systems or your edge cases. When it gets deployed for real, the solutions engineer has moved to the next prospect, and whatever breaks is now your problem or your implementation partner's.

This is not a criticism of the role. It is what the role is for. The mistake is expecting a pre-sales proof to be a deployment.

What is the difference between a solutions engineer and a sales engineer?

In most companies, nothing much. The titles are used interchangeably. Where there is a difference, it is this: a sales engineer sits in the sales organization and carries a number. Their compensation depends on deals closing. A solutions engineer may sit in a separate pre-sales or customer engineering group and be measured on deals supported rather than deals won.

From your side of the table, the distinction that matters is the incentive. A person whose bonus depends on you signing is not the person to ask whether you should sign. Ask them how it works. Ask someone else whether it is the right thing to buy.

What does a solutions architect do?

A solutions architect designs the system. They take your requirements and your existing landscape and produce the design: which components, how they connect, where the data lives, what the integration points are, what the risks are. In a large project the architect is essential, because someone has to hold the whole picture while different teams build the pieces.

I have done this job for a long time, and here is its weakness. The architect usually does not write the production code, and often is not there when the design meets reality. A design that looked correct on the whiteboard can fail on the second week of implementation because one system's API doesn't do what its documentation says, or because the data is dirtier than the requirements assumed. If the architect has moved on, the builders improvise, and the design and the system drift apart.

On a small engagement the same person designs and builds, and the problem goes away. That is most of what I mean by forward deployed: the architect stays, writes the code, and finds out whether the design was right.

Where a forward deployed engineer fits

A forward deployed engineer works after the sale, or with no sale at all. They are inside your company, using your real data, subject to your rules, and they are responsible for the system running in production. Not for a proof of concept, not for a design document. For the thing working.

In practice that means the engineer does parts of all three other jobs. Some of the solutions engineer's work, because the first weeks are spent proving what will and won't work on your data. Some of the architect's, because someone has to design it. And then the part none of the others do: building it in your environment, hitting the limits of your platforms and vendors, fixing the causes rather than the symptoms, and handing it over so your team can run it. I have written up how that runs week by week, from the records of a real engagement, if you want the detail.

The cost is that this person is expensive per hour and is with you for months. The return is that the money you spend produces something that runs, which, if you have been through an AI pilot that impressed everyone and shipped nothing, is the only return that matters.

Forward deployed engineer vs software engineer

This is the comparison engineers search for, usually because they are considering the job. The short version: the code skills overlap almost completely, and the job is different.

A software engineer builds a product for many customers, inside the vendor, against a roadmap. The constraints come from the product plan. A forward deployed engineer builds for one customer, inside that customer's business, and the constraints come from that company: its data, its approval process, the platforms it publishes to, the founder's rules. A software engineer can say "that's not in scope for this release". A forward deployed engineer is standing next to the person who needs it by Friday.

There is a temperament difference too. Forward deployed work rewards people who like finding out what the real problem is more than they like building the elegant solution to the stated one. If you are the kind of engineer who is bored by messy data and other people's rules, you will hate it. If you are the kind who is bored by building the same feature for a roadmap, it is the best job in the industry right now.

Forward deployed engineer vs AI engineer

An AI engineer builds and tunes models and the systems immediately around them: training pipelines, evaluation, serving. They usually work inside a product team or a research group. A forward deployed engineer takes models that already exist and makes them correct inside one company: what the model may read, what it must never say, what happens when it is wrong, who checks it, and how it gets into the systems staff already use.

Today most forward deployed work is AI work, because that is where companies have a capability they can't connect to their business. But the forward deployed engineer almost never trains a model. On the engagements I have run, the models were the part that took the least time. Everything around them took the months.

Which one should you hire?

• You are deciding whether a product fits. You want the vendor's solutions engineer, and you should not be paying for them. Ask for a proof of concept on your data, not theirs.
• You have a design and a team that can build it. A solutions architect, for the design phase, with a plan for who owns the design once building starts.
• You have a problem inside your own process that no product solves off the shelf, and you need it running, not recommended. A forward deployed engineer. Most mid-sized companies that have had an AI pilot fail are in this case, whether or not they have the name for it.
• You are not sure which case you are in. That is normal. A short advisory piece to find out is cheaper than hiring the wrong one of the three.

Common questions