The GTM engineer library.
What the role actually is, measured from 580 job descriptions, plus every asset worth reading, in one place.
A GTM engineer builds the systems that run go to market. Not the campaigns themselves, the machinery underneath them: the data pipelines that decide who gets contacted, the triggers that fire when something changes, the agents that draft and route, and the measurement that proves any of it worked. The job sits between marketing, sales and engineering, and it exists because the tooling finally got good enough that one person can run what used to need a team.
Most definitions of this role are opinion. This one is counted. We pulled 580 live job descriptions and tallied what employers actually ask for, so the numbers below come from postings rather than from a survey or a guess.
What GTM engineer jobs actually ask for
Counted from 580 live job descriptions, not from opinion.
Tools they name
Clay now outranks both CRMs in these descriptions.
What you have to be able to build
More than half ask for API work. This is the line between a GTM engineer and a GTM operator.
AI, specifically
A third want agents, not chat.
The work itself
Experimentation beats reporting. Barely.
Start here
The two assets we maintain ourselves, both free and both updated.
Every company hiring a GTM engineer right now 371 live roles at 334 companies, with the salary each employer posted and a direct link to the application. Refreshed weekly. 16 GTM plays with receipts Six plays we run at Earleads and ten from published teardowns. Each carries its trigger, its stack and the number its author reported.Understand the discipline
What the job is, how it differs from the roles it replaced, and what the first months look like.
What an agentic GTM operating system actually is GTM engineer versus an SDR team The four hats a GTM AI engineer wears Your first 30 days as a GTM engineer Agency versus fractional GTM operatorLearn the tooling
The primary documentation. Read the source before the summaries.
Claude Code documentationexternal Anthropic. The agent most of this work now runs inside. Model Context Protocolexternal The open standard for connecting an agent to your tools. Worth understanding properly. Clay documentationexternal Named in more of the job descriptions we read than either CRM. Build your own GTM agent Claude Code skills, and what they are forDocumented plays, with published numbers
Campaigns somebody actually published a result for. Vendor-reported, so treat the numbers as directional.
Common Room on job change playsexternal Reports 17% reply and 10% booked meetings, roughly 2x their cold baseline. Common Room on de-anonymised website visitorsexternal Person-level resolution tops out near 20% of traffic. Work the account for the rest. GTMinds on 50+ Clay deploymentsexternal Single-provider email coverage runs near 50%. A waterfall takes the same list past 80%. Outbound Republic case studiesexternal Campaigns with real denominators, which is rarer than it should be. Unify on building a signal based playbookexternalPick your lane and go deep
Generalists get screened out in round one. These are the lanes worth owning.
Signal based outbound and buying triggers AI agents for outbound campaigns AI agents for RevOps Waterfall enrichment, and why one provider is not enough Turning website visitors into outboundKnow the stack
Our reviews and directories. Opinionated, and we say when a tool is the wrong choice.
The GTM tools directory MCP servers worth wiring in Runnable Claude Code skills What a modern GTM stack looks like Clay alternatives, honestly comparedIf you want the job
Pick one lane. Build one system in it end to end, including the measurement. Write the hypothesis down before you start so the result can prove you wrong. Then take the number you got into the conversation. A candidate who can show what their system changed is much rarer than one who can show what they built.
Want this built into your own stack?
We build these systems for a living. Bring the channel you want to own and we will map what it takes.