what is a gtm engineer

At the installation company whose systems I run, an enquiry comes in and a whole lot happens that nobody looks at anymore. The lead lands in the CRM. A call gets scheduled. The quote goes out, the signed agreement comes back, the deposit invoice gets raised, the work order follows, and afterwards the customer automatically gets asked how it went. I built that whole path and I keep it running. Just my job, I figured.

Turns out that job now has a title in America. GTM engineer. Go-to-market engineer, in full. And there’s an entire circus attached: newsletters, communities, salary bands, people arguing on LinkedIn about what it even is. So I went digging. If my work suddenly has a name, I’d like to know what they mean by it.

where the term suddenly came from

Short answer: a software company. Clay, an American data vendor, couldn’t fill its enterprise sales vacancy back in 2023. One of the founders rewrote the job ad, called the role GTM engineer, and the name stuck. First internally, then in their marketing, and by now in thousands of vacancies worldwide. From nearly zero to over three thousand open roles in about two years; earlier this year an analysis of job-board data showed the count doubling year on year.

That’s real growth, not vapour. But keep in mind who coined the word. A company selling tools for this work has every reason to want the work to become a profession. And the profession is young enough that nobody gives you the same answer twice: in a benchmark among practitioners themselves, less than half of organisations say they clearly understand what their GTM engineer actually does. Honestly, that’s my favourite number in the whole pile. A role growing this fast while half the employers can’t explain it.

Pays fine, mind you. The median in US postings sits around 127,500 dollars. You’ll see bigger numbers floating around, but those come from the top band, not the middle.

what a gtm engineer actually does

Underneath the buzzwords sits something very recognisable. A GTM engineer builds the machinery between “someone shows interest” and “the money is in the account”. Enriching lists of potential customers with data. Catching signals and acting on them automatically. Wiring the CRM to email, calendar, invoicing. And lately: AI agents doing chunks of that work on their own. The usual line between this and RevOps, the older sibling: RevOps runs what exists, the GTM engineer builds what doesn’t. In reality the two blur constantly, and part of the field thinks the whole distinction is marketing.

The toolkit is strikingly uniform. Clay itself, obviously, workflow tools like n8n, a CRM, and Claude as the AI model that gets named most. I’ll admit I grinned reading that list. My own stack for that installation company: monday as the CRM, Make for the workflows, Claude with MCP connections alongside. Different logos, same machine. I even built a dashboard next to it that pulls the revenue side out of three separate systems, with response speed on new enquiries as the headline gauge. In America that’s a revenue dashboard and it belongs to GTM engineering. Here it’s called knowing how your business is doing.

And it’s not just an American thing anymore. The vacancies are showing up here too; a software company in Rotterdam recently advertised for an “AI GTM Engineer (RevOps)”, with HubSpot, n8n, Clay and Claude with MCP literally in the job text. Everything I just described, in one Dutch vacancy.

why you probably shouldn’t hire one

Now the part where I have an opinion not everyone will like.

If you run an SMB, say 5 to 50 people, odds are you’ll never put a GTM engineer on payroll. Rightly so. The profile is scarce, expensive, and built for companies with a sales team that needs feeding. That’s not you. But here’s the thing: the work behind the title? You do need that. Every enquiry that sits for days, every quote nobody chases, every invoice that only goes out when someone remembers: that’s exactly the gap this field exists to close. And that gap costs you revenue just as surely as it costs a Rotterdam scale-up. The scale-up hires someone for it. You just keep living with it.

The good news is the maths has changed. This work used to take a team: a developer for the integrations, a marketer for the follow-up, someone for the reporting. With today’s AI tooling, one person builds the whole chain, and then it runs. I do this work myself for that installation company, from lead to satisfaction email, and I’m not a team. So what you need isn’t a hire. It’s the chain, built properly once, and a system you can talk to instead of click through.

As for the title itself? I give it a few years. Analysts already expect the label to dissolve into RevOps or something called AI Ops. Fine. Titles blow over, they always do. The work, signal to system to revenue, doesn’t blow over. How you organise that work without standing up a department is what the next essay is about.