What Is a GTM Engineer?

Clay coined the role in 2023 and defines it as building automated revenue systems out of AI, data enrichment and workflow automation. Two years later it is the fastest-growing job in go-to-market, and it is priced like software engineering.

+205%GTM engineer and RevOps postings
Jan to Sep, 2024 vs 2025

~100GTM engineer listings going
live monthly, per Clay

$127.5Kmedian salary, $250K+
at Vercel and OpenAI

Bloomberry's analysis of more than 1,000 GTM engineering listings was published on 3 October 2025 and is explicit about scope: its 205% counts GTM engineer and RevOps postings together, January through September in each year, and it puts the median at $127,500 on listings read through September 2025. Clay, which coined the title, counts about 100 of those listings going live every month. eMarketer, on 31 March 2026, reports Vercel at around $252,000 and OpenAI at around $250,000 and credits Bloomberry for both. A marketing hire priced like a senior software engineer, because it is one.

What does a GTM engineer do?

A GTM engineer ships systems, not campaigns. Six outputs cover most of the job.

Signal enginesScrapers + enrichment that find in-market accounts before they raise a hand

AgentsCompetitive intel, campaign QA, research agents running on schedule

Landing systemsProgrammatic pages and full features shipped without an eng ticket

Data pipelinesAttribution, incrementality, CRM hygiene wired end to end

Outbound machinesScoring, routing and personalized sequences from live data

Reporting brainsDashboards agents act on, maintained like a codebase

The mix shifts by company. An early-stage team wants the outbound machine first. An enterprise usually wants the pipelines and the CRM cleaned before anything else gets built on top. See six example GTM engineer profiles for what the shipped work and the stack behind it look like in practice.

The postings describe the same six outputs in their own words. AssemblyAI asks the hire to "own the GTM systems stack: Salesforce, outbound and sequencing tools, enrichment providers, and the integrations between them". Together AI wants someone to "identify repetitive work across the GTM org" and replace it "with intelligent automation using APIs, webhooks, and AI agents". Cresta describes the output as "systems that tell sellers what to focus on each day and what actions move deals forward".

Three companies, three phrasings, one job: own the stack, remove the repetitive work, and turn data into something a seller acts on before lunch. Note what none of them ask for. Not a campaign calendar, not a content plan, not a quota. The unit of work is a system that keeps running after the person who built it goes to a meeting.

What does GTM stand for in engineering?

GTM stands for go-to-market. In this title it does not name a branch of engineering the way backend or platform engineering does. It names the department the engineer builds for.

That distinction is most of the job description. A product engineer builds the thing the company sells. A GTM engineer builds the machine that sells it: the pipeline that finds accounts, the enrichment that scores them, the routing that assigns them and the reporting that shows which half worked. Same craft, different customer. Here the customer is the revenue team.

Two things follow from that. The work is judged on pipeline and revenue rather than uptime and latency, so an engineer who cannot read a funnel is missing half of it. And the seat usually reports into revenue or marketing rather than into engineering, which is why the same person is called a marketing engineer at one company and a GTM engineer at the next.

GTM engineer vs RevOps vs marketing ops

They own different things. RevOps owns process, governance and cross-functional alignment, keeping the existing motion reliable and consistent. The GTM engineer builds new engines on top of it. Marketing ops sits closest to the campaign layer and the martech administration around it.

GTM engineer

Builds new engines. Signal engines, agents, pipelines and outbound machines that did not exist last quarter. Measured on what shipped and the number attached to it.

RevOps

Keeps the motion reliable. Process, governance, forecasting and cross-functional alignment. Measured on whether the existing machine runs consistently.

Marketing ops

Runs the campaign layer. Martech administration, lifecycle plumbing, lists, attribution reporting inside the marketing stack. Measured on campaign throughput and data quality.

In practice one person often holds two of these seats at a smaller company. The distinction still matters at hiring time, because a RevOps job description handed to a GTM engineer produces someone maintaining dashboards instead of building engines.

What is a typical GTM engineer salary?

Two medians get quoted for this role and they are $22,500 apart. Bloomberry's analysis of more than 1,000 listings, published on 3 October 2025, puts it at $127,500. Steven Moody's benchmark of 235 priced US roles, published May 2026, puts the median base at $150,000. Neither is wrong. The title repriced in the eight months between them while the older figure was still being copied between articles.

$127.5Kmedian across 1,000+ listings,
Bloomberry, October 2025

$150Kmedian base, 235 priced US
roles, Steven Moody, May 2026

$150k–$230kwidest band posted by one
employer, read 7 Sep 2026

eMarketer, on 31 March 2026, puts Vercel at around $252,000 and OpenAI at around $250,000 for the title, citing the same October 2025 Bloomberry analysis. The spread is wide because the job is two jobs: the judgment of a growth marketer and the hands of an engineer, in one head. Companies paying at the top are pricing against senior software engineering, not against marketing.

The deeper reason the numbers scatter is that GTM engineer is not one job. Steven Moody's benchmark of 235 priced US roles sorted them by what the posting actually asked for, and the pay separates cleanly by archetype.

That is an $83,750 spread inside one job title, and the two biggest buckets are the two cheapest. Before writing a band, decide which of those five you are actually hiring. A posting that describes the $135,000 archetype and pays the $218,750 one wastes money; the reverse wastes a quarter.

Published medians also age, and a band an employer is advertising today does not. These four were live on 7 September 2026.

Two things to take from that table. Every one of those floors sits above the $127,500 median still being quoted as the going rate, which is the clearest sign the older figure trails the market rather than describing it. And the bands are not measuring the same quantity: Together AI posts base only, SecurityScorecard posts base plus bonus, Cresta attaches equity it does not price. Comparing a base-only floor against a base-plus-bonus ceiling produces a spread that is partly an artifact of what each company chose to publish.

Seniority and location move the number about as much as you would expect, and less than most people assume. On the same 235-role benchmark the median base runs $150,750 at mid-level, $170,000 at senior and $200,950 at staff. By location it is $171,250 in the Bay Area, $150,000 in New York, $150,000 for a remote US role and $145,000 elsewhere in the US. A remote role pays the New York median. The penalty for not being in San Francisco is roughly $21,000, which is smaller than the gap between two archetypes sitting at desks in the same city.

One finding in that benchmark deserves its own line. Roles whose postings mention Clay averaged $148,174, against $165,000 for roles that do not, a $17,000 discount for naming the tool. Naming a specific platform tends to describe an operator who runs it rather than an engineer who builds around it, and the market prices the difference.

If you are pricing the role rather than applying for it, the comparison is a salary plus ramp against someone who arrives already able to build. The monthly rate for what an embedded GTM engineer costs is published, so you can put the two numbers side by side before you write a job description.

Why GTM engineer salary numbers disagree

Search the term and most of page one quotes the same two figures, $127,500 and about $250,000 at Vercel, and nearly all of them trace back to one Bloomberry analysis published on 3 October 2025. Aggregated numbers get copied between articles faster than anyone re-measures them, and the copy usually drops the date, so a median measured on listings read through September 2025 reads as a 2026 one.

The check that settles it is to open the employer boards and read what is posted today, under pay transparency rules that oblige the employer to honor the band. That check has a cost worth stating. Of roughly ten indexed GTM engineer postings opened on 7 September 2026, four were still live; the rest had been taken down or had redirected to a careers index. A posted band is the most reliable number available and the shortest-lived, which is why every band on this page carries the date it was read.

There is a deeper reason the numbers scatter, and it is worth knowing before you write a job description. The title came out of a vendor ecosystem rather than an HR job architecture: Clay coined it in 2023 and, in its own account, says that “since we coined this role in 2023, it has emerged at companies like Cursor, Lovable, and Webflow.” Nobody standardised it afterwards. So two companies advertising a GTM engineer are not reliably advertising the same job, and the responsible move when comparing bands is to read the responsibilities and discount the title.

What skills does a GTM engineer need?

Two things at once: the judgment to pick the right motion and the hands to ship it. Read how the best companies write the job and the profile gets specific fast.

Ramp hires an "Agentic Operator" onto its growth marketing team, and the listing reads like a spec. "Bias toward shipping over spec'ing. You believe a working prototype today is better than a polished deck next week." And, "think like a growth marketer but move like a hacker."

• Workflow and agent tooling. The Ramp listing names GPTs, Python notebooks and Make workflows. One tool deeply beats five shallowly.
• APIs and webhooks. The plumbing under every GTM stack. Enough to wire two systems together without asking engineering.
• Data modeling and CRM hygiene. Deduplication, field discipline, the written definitions of account and opportunity. Dirty data breaks agents faster than bad prompts.
• Go-to-market judgment. Which motion to build first, and which number it moves. This is the half that cannot be picked up from documentation.

What tools does a GTM engineer actually use?

Most published skill lists for this role name the same generic four: SQL, Python, a CRM and a workflow tool. The live postings give a sharper answer, because an employer naming a product has usually already bought it.

Across the four postings priced above, read on 7 September 2026, two products appear in every one.

• Salesforce, four of four. Not as an administrator. SecurityScorecard's posting asks for Apex, Flow, custom objects, REST APIs, Git, Postman, the Salesforce CLI and Gearset, which is a developer's relationship with the platform rather than an admin's.
• Clay, four of four. The enrichment layer has effectively standardized. It is also the one product whose presence in a posting predicts a lower band, for the reason in the section above. A posting tells you the tool is in the building; it does not tell you what a Clay GTM engineer does with it all day, which is where the tables, the waterfalls and the agent layer come in.
• An Anthropic coding tool, three of four. SecurityScorecard and AssemblyAI both name Claude Code; Together AI names Claude Cowork. The AI tool in these postings is one the engineer builds with, not one the company is buying to replace them.
• A warehouse and a modeling layer, two of four. Together AI names Snowflake, AssemblyAI names BigQuery, dbt and Looker. This is the archetype the benchmark prices at $166,195.
• Revenue tooling around the edges. Gong, Clari, ZoomInfo, 6sense, LeanData, Gainsight, Workato, Zapier and Slack all appear, none of them in more than half the postings. The edge of the stack is where companies differ.

Two cautions on that list. Four postings is a small sample, and it is skewed toward venture-funded software companies hiring in San Francisco, which is where this title is concentrated. What it is good evidence for is the shape of the ask rather than the market share of any product.

The pattern underneath is that these companies want someone who writes code against a CRM rather than someone who configures one. That single distinction separates this role from marketing ops more reliably than any of the definitions do, and it is the thing to test for in an interview.

How do you become a GTM engineer?

Ship one system end to end and attach a number to it. The role is new enough that a working portfolio outranks a resume, and the people getting hired are growth marketers who developed what Jonathan Martinez calls the itch to build, the reflex that asks how to automate something before asking who to brief.

• Ship one system end to end. A signal scraper, an enrichment flow, a programmatic page set. Small is fine, running is mandatory.
• Learn the plumbing under your stack. APIs, webhooks, one workflow tool deeply.
• Attach a number to everything. A system with a before-and-after metric is a portfolio. A certificate is not.
• Put the working demo first. The prototype-today culture in Ramp's listing is how the whole category now interviews.

The postings agree on the entry bar more closely than the salary data agrees on anything. SecurityScorecard asks for three or more years in GTM engineering, Salesforce, business systems or revenue operations. Together AI asks for three to five or more in a growth, GTM engineering or RevOps role. AssemblyAI asks for four or more in GTM or sales engineering, growth engineering, RevOps, or software engineering with GTM exposure.

Every one of those lists names RevOps as a qualifying background, which makes it the most reliable side door into the title. It also means nobody in this sample is hiring at zero years. That is the honest version of the argument that a new role has no seniority gate: the ceiling is not gated by tenure, but the floor still is.

Read the archetypes from the other end and the career path is legible. The outbound and growth archetype starts at a $135,000 median and the software archetype at $218,750, and the distance between them is the half of the job you are missing. An operator closes it by learning to ship. An engineer closes it by learning to read a funnel. Both routes end at the same seat, and the second one is rarer, which is why the org chart below treats the engineer seat as the hard one to fill.

Is GTM engineering a good career?

The demand data is strong and the supply is thin, which is the combination that moves salaries. Postings for the GTM engineer and RevOps roles, counted together, grew 205% year over year on Bloomberry's count, and Clay puts roughly 100 new GTM engineer listings on the market every month.

The scarce side is people who can hold a pipeline target in their head while shipping the system that moves it. Most growth marketers can do the first half. Most engineers can do the second. The overlap is small, and the role is young enough that nobody holds ten years in it, so the ceiling is not gated by tenure the way senior marketing roles usually are.

The posting counts back up the salary movement. Stacker, reporting in January 2026, counted more than 1,400 GTM engineering postings on LinkedIn in mid-2025 and more than 3,000 that January, and cites ZoomInfo data showing hiring for the role doubling year over year for two consecutive years, with peaks in January and July. If you are timing a move, those are the two months when the listings appear.

One caution against all of it. This is a young title with a short measurement history, and almost everything published about it counts job postings rather than filled seats. A posting boom tells you what companies are trying to buy. It does not tell you how many of those seats were filled, or how many people were still sitting in them a year later, and that data does not exist for this role yet. Anyone quoting you a retention or seniority statistic for GTM engineers is estimating.

Where the GTM engineer sits: the three-seat org chart

Inside a repeatable unit of three seats, dropped into each go-to-market function. I build these org charts inside companies, so here is the actual shape rather than a gesture at "pods."

• The operator. Owns the business problem and the outcome. Usually internal.
• The ops lead. Owns the systems, the data and the definitions. Usually internal.
• The GTM engineer. Builds. This is the hard seat to fill.

The engineer seat has two shapes. One engineer can serve marketing plus sales, or all three functions, dotted-line into each team. Most companies start there and split into dedicated seats per function once the build backlog justifies it.

Above that sits one thin central layer, a hub, that owns the shared brain and the standards. Keep it purely enabling, every function pulls from it, and do not let it turn into a center of excellence that hoards the work. The hub sets the standard, the departments ship. One brain underneath all of them.

The hubShared brain + standards. Thin, purely enabling. Every function pulls from it.

Marketing

OperatorOwns the problem + the outcome

Ops leadOwns systems, data, definitions

Sales

OperatorOwns the problem + the outcome

Ops leadOwns systems, data, definitions

Customer success

OperatorOwns the problem + the outcome

Ops leadOwns systems, data, definitions

GTM engineer(s)Builds. The hard seat to fill. Dedicated per function, or one engineer dotted-line across the functions.

Setup 1 · SharedOne engineer across marketing + sales, or all three. Where most companies start.

Setup 2 · DedicatedOne engineer inside each function, once every function has a full build backlog.

Marketing is the function this usually hits first. It runs on repeatable text-and-data work, enrichment and scoring and QA and versioning and list building, which is exactly what agents do well. That is also why the same seat gets called a marketing engineer inside some companies and a GTM engineer inside others.

GTM engineer, growth engineer, marketing engineer: same role?

Close to it. The three titles describe the same seat approached from three different starting points, and the job descriptions have converged on the same output.

GTM engineer comes from the sales and revenue side, coined at Clay in 2023, and it is the title the published market data measures. Growth engineer comes from growth marketing; Martinez frames it as thinking less about running experiments and more about building the infrastructure that enables them. Marketing engineer is the version that shows up when the seat is created inside a marketing org rather than a revenue org, and it is usually the same job description with a different reporting line.

Two practical consequences. If you are hiring, search all three titles or you will miss most of the market. If you are applying, GTM engineer is the title with the salary benchmarks attached, so it is the one to put on the resume.

Is a GTM engineer the same as a forward deployed engineer?

No, and the difference is whose building you work in. A forward deployed engineer embeds inside a customer's company and builds their own employer's product into that customer's environment until it runs in production. A GTM engineer builds inside their own company, for their own revenue team. One is pointed outward at an account, the other inward at a funnel.

The confusion is fair, because the working style is nearly identical. Both sit next to the person with the problem instead of taking a ticket. Both are measured on something running in production rather than on a document. Both were created because the gap between what a team needed and what a central engineering queue could deliver had become the bottleneck. Palantir reached that conclusion for customer deployments a decade before Clay reached it for revenue teams.

The two titles meet in a third one. A forward deployed GTM engineer is an engineer embedded in somebody else's revenue organization, building their go-to-market systems against their real CRM and their real data, and staying until those systems run. The production system is the pipeline, and the measure is pipeline. That is the shape of the borrow route below, and it is worth naming precisely because almost nobody is hiring under that exact title yet, so there is no published salary data for it. Any number quoted for that hybrid today is an estimate rather than a market rate.

Proof it works

Two companies have published their numbers, one at scale and one inside a marketing org.

Uber embedded the builders. Praveen Neppalli Naga, Uber's CTO, described pairing around 30 AI-proficient engineers with domain experts in two-week agentic pods, and 16 of them ran across 16 business functions in two months. Each pod shadows the real workflow first, builds against real systems with the person who does the job, and ships on day ten.

01PairOne engineer, one domain expert, one pod

02UnderstandShadow the real workflow for two days before building

03IdentifyRank the opportunities by scale, repetition, impact

04BuildAgents and workflows against real systems, days four and five

05ValidateTest with others who do the same job, measure the impact

06ShipDay ten. In production, in the hands of the team.

The published results: marketing QA on campaigns went from two weeks to under an hour, capital allocation analysis from 15 hours to 30 minutes, financial pacing reports from two days to ten minutes. The backdrop makes it possible. By Uber's own count 99% of its engineers use AI tools and more than 70% of pull requests are attributed to local or cloud agents.

2wks → 1hrmarketing QA
after one pod

70%+of pull requests
attributed to agents

16 podsacross 16 functions
in two months

Read that pod structure again and the three seats are already in it. One operator who owns the outcome, one person who knows the systems, one engineer who builds, discovered independently by a company with tens of thousands of employees and a share price.

Backbase is the pure marketing version. Its CMO Tim Rutten published the numbers himself after the €2.5B digital banking platform rebuilt its go-to-market around in-house builders. Account coverage per ABM marketer up 10x. New and qualified pipeline up more than 85% year on year, average deal size up more than 60%, on a marketing budget optimized by 25%. The ABM platform, the integration stack and the reporting stack were all absorbed into something the team built itself. Thirty-five people run a $350M+ ARR book on it, and the signal engine scans 3,000 accounts for about $500 a week.

Neither company bought its way there with software. Both changed who sits inside the go-to-market org and how the work ships.

How to interview a GTM engineer

The archetype problem shows up hardest in the interview, because all five archetypes describe their work with the same vocabulary. Everyone says automation, everyone says pipeline, everyone says they built something. Four questions separate them.

• Show me a system that is still running. Not a screenshot, not a deck. Who maintains it now, what broke last, and how did they find out it broke? An operator demos an output. An engineer talks about failure modes and monitoring without being asked.
• What number did it move, and how do you know it was you? The second half is the real question. Someone who has shipped inside a revenue org has been asked to defend attribution and has an answer ready, including the parts they cannot prove.
• Here is a messy CRM export. What would you fix first? The strong answer starts with definitions, what counts as an account and what counts as an opportunity, before it touches deduplication. Dirty data breaks agents faster than bad prompts do, and the people who have been burned know it.
• What would you deliberately leave manual? The most useful signal on the list. Anyone who wants to automate everything has not yet maintained anything. Judgment about what not to build is the thing that separates a $218,750 hire from a $135,000 one, and it is the only one of the four you cannot coach.

Skip the take-home. This role gets hired against a portfolio of running systems, and the good candidates have one already; asking them to build a throwaway is how you lose them to the company that just asked to see what they had.

How to hire a GTM engineer

Three routes, and the market rate sets the floor under all of them.

• Hire. Budget against the $150,000 median base from the May 2026 benchmark rather than the older $127,500 figure, which every band in the table above already clears. Expect to lose candidates to the companies paying $250,000, and search growth engineer and marketing engineer as well as GTM engineer.
• Borrow. An embedded engineer who has built this before, on a monthly rate, with no ramp and no hiring risk. Fastest route to a shipped system.
• Grow. Take the operator on your team with the itch to build and give them one system to own. Slowest, and the only route that compounds internally.

Whichever route you pick, start with one workflow rather than a program. I have installed this unit more than 20 times, across B2B and B2C, from seed-stage startups to a €2.5B enterprise, and the first build is always one painful repeatable process with a number attached.

One correction to make before the budget goes anywhere near a job description. Decide which archetype you need first, because the five of them are $83,750 apart and they are not interchangeable. A team whose CRM is already clean and whose problem is that nothing gets built needs the software archetype and should budget past $200,000. A team drowning in manual list work needs the outbound archetype at $135,000 and would waste the other hire. The most common failure is writing a RevOps job description, paying an operator band, and expecting engines to come out. The rest of that decision, the readiness test, what the search costs in time and the questions that separate a builder from a title, is in the buyer's guide to hiring a GTM engineer.

01Pick the worst workflowOne painful, repeatable process with a number attached

→

02Run one podOne engineer + one operator, shadow first, ship in two weeks

→

03Publish the numberBefore and after, internally. Let the org chase the evidence.

What the engineer builds toward is a shared system, not a pile of automations. I recorded a walkthrough of a working GTM OS: signals, workflows, agents and reporting in one place, operated by a three-seat unit.

Questions people ask about the role