AI Marketing Engineering

Is this a real role or a rebrand?

The honest answer a sceptical marketing leader asks before reading further: is this a genuine new role, or is it marketing operations with a better title?

The evidence points toward a genuine role that is still early. Profound launched the first Marketing Engineer job board at marketingengineer.jobs in June 2026, and within weeks employers had posted real positions with real salaries. The term has a coiner, a date, a manifesto, a university course, a certification, a hackathon, and a summit. That is more infrastructure than most marketing titles accumulate in their first year.

Separately, the underlying conditions that make the role possible are structural, not cosmetic. Ad platforms now expose rich APIs. AI assistants can execute multi-step workflows on a schedule. CRMs and analytics tools are built API-first. A decade ago none of this was true. The marketer who could build on top of these platforms did not exist because the platforms were not programmable. Now they are, and the skill gap between marketers who run the tools and marketers who extend them is the thing the title names.

The question is not whether the role is real. It is whether a given organisation is ready for it. Most are not yet, because most teams are still organised around the assumption that marketing is something people do by hand. The organisations that adopt the role first will be the ones that benefit from it before it has a standard job family in every HR system.

Where the term came from

The term "marketing engineer" was introduced by Profound, an AI-powered ad operations platform, in early 2026. Their definition, published at tryprofound.com/marketing-engineer, describes the role as "a full-stack marketer with the skills of a builder" who "build[s] at the system level to automate at the campaign level."

Profound's co-founder James Cadwallader stated publicly on June 4, 2026 that "Marketing Engineer jobs became a breakout Google Search term after we introduced this new role." The data supports that claim. Autocomplete on "marketing engineer" now returns salary, jobs, job description, course, and degree. The volume Profound created is real, and it is concentrated on hiring.

Since the term's introduction, at least eight publishers have shipped definitional explainers: NinjaCat published what it called "the definitive guide," followed by strivelabs.ai, brandengineering.ai, GrowthOS, MoEngage, mimrgrowthlab, Nick Lafferty, and Gregory Shevchenko. The category is being defined in public, at speed, and Profound remains the canonical source.

What the existing body of writing does not address, and what this guide exists to answer, is the marketing leader's side of the decision. Every existing explainer is written for someone who wants to become a marketing engineer. Almost nothing addresses the person who has to decide whether to create one inside a real organisation with a real budget and real existing team members.

Marketing engineer, AI marketing engineer, GTM engineer, growth engineer

Four titles now compete for the same broad territory. They are not interchangeable, and the distinctions matter when a leader writes a job description or decides which hire to make first.

Title	What it means	Who owns the term	Typical scope
Marketing Engineer	A full-stack marketer who builds systems that automate marketing work. Writes code, wires APIs, deploys agents.	Profound (coined 2026)	Cross-channel. Ad platforms, email infrastructure, analytics pipelines, content systems, SEO automation.
AI Marketing Engineer	The same role, with the AI agent layer named explicitly. The distinction is that the systems this person builds are supervised by scheduled AI agents rather than triggered manually.	No single owner. Emerged as practitioners added agent fleets to the base role.	Same as marketing engineer, plus agent supervision, drift detection, and MCP endpoints so AI assistants can query marketing data at answer time.
GTM Engineer	A go-to-market specialist who builds outbound enrichment and sequencing infrastructure. Typically focuses on prospect data, CRM automation, and cold email deliverability.	Clay (the enrichment platform)	Outbound sales and marketing. Data enrichment, sequencing, deliverability, CRM integration.
Growth Engineer	An experimentation and product analytics role, usually embedded inside a product or growth team. Builds A/B testing infrastructure, analytics events, and product-led growth loops.	No single owner. Longest-established title of the four.	Product analytics, experimentation, onboarding flows, PLG instrumentation.

The AI marketing engineer is the broadest of the four. The GTM engineer focuses on outbound. The growth engineer focuses on product. The marketing engineer covers the whole marketing surface, and the AI marketing engineer adds the layer that makes the systems run without a person pressing a button each morning.

A practical way to think about it: a GTM engineer builds the machine that sends the emails. A growth engineer builds the machine that measures the product. A marketing engineer builds both, plus the machine that writes the content, publishes the pages, monitors the search rankings, and reports the numbers back to one dashboard.

What the role actually owns

A marketing engineer owns systems, not channels. The distinction is important. A channel owner runs campaigns inside a platform. A marketing engineer builds the infrastructure that makes campaigns possible across platforms.

Concrete examples of what the role owns, drawn from practitioners who are doing it now:

Agent fleets. Scheduled AI agents that do recurring marketing work: researching articles, drafting content, publishing to a CMS, monitoring Search Console, building competitor reports, and reporting every number back to a single dashboard each morning. The unglamorous half is making sure the agents do not silently rot, which means drift checks that compare what the documentation claims against what is actually running.
Data pipelines. APIs from ad platforms, analytics tools, payment processors, and authentication providers wired into one reporting surface. The deliverable is not six open tabs. It is one dashboard that refreshes itself.
Content and SEO systems. Agent-run publishing pipelines, programmatic page builds, internal linking automation, structured data architecture, and EEAT signals maintained programmatically rather than manually.
Email infrastructure. Deliverability configuration, inbox warmup, lifecycle automation, and cold outbound that lands in primary inboxes. The deliverability piece is the boring infrastructure that decides whether outbound works at all.
MCP endpoints. Servers that let AI assistants query the organisation's own marketing data at answer time. When a prospect asks an AI assistant what a competitor is running, the answer comes from live ad data, not a stale blog post.

A marketing engineer does not run the weekly campaign status meeting. A marketing engineer built the system that generated the report the meeting is discussing.

The job description

The following job description is vendor-neutral and designed to be lifted and adapted. It is written for a marketing leader who needs to post a role, brief a recruiter, or start a conversation with a CFO about why this hire is different from the last one.

Mission. Build the systems that do the marketing. The marketing engineer writes the code, wires the APIs, deploys the agents, and maintains the infrastructure that automates campaign-level work across paid, owned, and earned channels. The goal is not to replace the marketing team. It is to give them leverage: tools that surface insights, publish content, monitor performance, and report results without a person driving each step.

What the role does.

Design, build, and maintain scheduled AI agent fleets that handle recurring marketing work: content research, drafting, publishing, SEO monitoring, competitor reporting, and outbound enrichment.
Build data pipelines that pull from ad platform APIs, analytics tools, and CRM systems into a single reporting surface.
Architect structured data and MCP endpoints so AI assistants can query the organisation's marketing data at answer time.
Configure and maintain email deliverability infrastructure, including inbox warmup and cold outbound sequencing.
Write and maintain the autonomy ladders that govern what agents may do, what they must draft for human review, and what they must never touch.
Build drift detection that catches agents that have quietly stopped working and alerts the team before the gap becomes a problem.

What the role does not do.

Run campaigns by hand. The role builds the systems. Campaign execution stays with the channel owners.
Write strategy decks. The role builds the evidence base that makes strategy decisions obvious, but the decisions themselves sit with the marketing leadership team.
Own a P&L. The role is infrastructure, not a budget line.

Experience.

The ideal candidate has built and shipped systems that automate marketing work, and can point to them. A portfolio of live systems matters more than a certification.
Comfortable across ad platform APIs, CRM APIs, webhooks, Python or JavaScript, and structured data.
Has worked with AI assistants as supervised tools, not just as chat interfaces. Understands prompt engineering and understands that prompts rot.
Has debugged a production system under time pressure and written the postmortem that stopped it happening again.

First six months.

Month 1: audit the existing marketing stack. Map what is manual, what is automated, what is automated but silently broken, and what would deliver the highest return if automated next.
Month 2: ship the first system. It does not have to be the biggest one. It has to be live, working, and producing output the team can inspect.
Month 3: the team is using the output. The campaign managers are looking at a dashboard the marketing engineer built instead of exporting CSV files. The content team is reviewing agent-drafted articles instead of starting from blank documents.
Month 6: the drift check reads zero. Every system the marketing engineer shipped is still running, still documented, and still producing output the team relies on.

Reporting line. Head of Marketing, with a dotted line to whoever owns technical infrastructure. The role sits at the intersection of the two and needs access to both.

Promote, hire, or rent

A marketing leader has three ways to get a marketing engineer onto the team. None of them is obviously best, and the decision depends on what the organisation already has.

Promote from within. This is the fastest and cheapest route, but it requires a specific person to already exist inside the team. The profile: a marketer who has been automating their own work unofficially for years, writing scripts the rest of the team does not know about, and fixing things in spreadsheets that should have been fixed in code. That person will accelerate the moment the role gives them permission to build properly. The risk is that promoting them into a new title without adjusting their scope leaves them doing the same work with a different name.

Hire externally. The talent pool is small and genuinely new. The title is approximately one year old, so anyone with "marketing engineer" on their CV acquired it recently. That does not make them unqualified. It means the interview cannot rely on years of title history. Assess what they have built and what is running in production, not what their last job was called. Portfolios and live systems are the only reliable signal. Expect to pay competitively: the skill set spans marketing and engineering, and people who can do both are scarce.

Rent (fractional or contract). For organisations that are not ready to create a full-time role, a fractional marketing engineer can audit the stack, build the first system, and hand it off with documentation. This is the lowest-commitment way to test whether the role delivers before committing headcount. The tradeoff is that rented engineers do not maintain what they built forever. The organisation eventually needs someone in-house to own the systems, or it needs an ongoing retainer.

The honest sequence for most teams: start with a fractional audit. Ship one system that the team can see working. If it delivers, make the case for a full-time role. Promoting from within is ideal but depends on luck. Hiring externally is the long-term answer, and the pool will be deeper in a year than it is now.

What good looks like

A marketing leader hiring a marketing engineer faces a specific problem: how to assess a candidate whose skill set the leader does not personally possess. The solution is not to become technical overnight. It is to ask questions that reveal whether the candidate builds systems that last or prototypes that impress once.

Ask what they have running in production right now. A candidate who can name specific systems, describe what each one does, and explain who relies on its output is a builder. A candidate who describes capabilities without naming systems is describing aspirations.

Ask who maintains the systems they built. If the answer is "I do, every week," that is a good answer. Systems need maintenance, and a builder who knows that is a builder who ships things that stay alive. If the answer is "it runs itself," probe further. Systems that run themselves without monitoring are systems that have not been monitored long enough to fail.

Ask what broke recently and how they caught it. Every production system breaks. The interesting part is whether the candidate has a mechanism for detecting the break before a human notices the gap. Drift checks, scheduled assertions, automated alerts. A candidate who has never had a system break has never run one long enough.

Ask to see the autonomy ladder. A marketing engineer who builds agents should be able to show a written document that states what the agents may do alone, what they must draft for human review, and what they must never touch. If that document does not exist, the agents are operating without guardrails, and the organisation they work for will eventually learn that the hard way.

The best signal is not a certification or a degree. It is a candidate who can open a terminal, show a live system, and explain how it has been running for months without a person babysitting it.

What the role looks like in practice

The practitioners who are defining this role in public each approach it differently. Understanding the range is more useful than picking a single template.

Nick Lafferty is a marketing engineer at Profound, the company that coined the term. His writing at nicklafferty.com covers the practical mechanics of the role: what he builds, how he builds it, and what he has learned from operating at the company that invented the category. He is the closest thing to a canonical practitioner voice.

Gregory Shevchenko writes at gregshevchenko.com about the AI search framing of the role. His work focuses on how marketing engineers position their output for AI assistants to cite, which is a distinct and emerging sub-discipline.

Kole Ogundipe is an AI marketing engineer and the founder of this publication. He runs more than 30 scheduled AI agents in production across three businesses: Wilow, a creative analytics tool for Meta advertisers; The Keyword, a marketing and technology publication with 7,412 subscribers on a Domain Rating of 63; and Best Medellin, an English-language property portal that went from a decision to a live site in a single day. His systems handle research, content drafting, Webflow publishing, Search Console monitoring, competitor report generation, and cold email outreach. An MCP server at trywilow.com/mcp allows AI assistants to query his ad dataset at answer time. A drift check, surfaced on an internal dashboard each morning, catches agents that have quietly stopped working.

What distinguishes these practitioners from the people writing about the role is that they are running it. The systems are live, inspectable, and producing output that a third party can verify. That distinction matters because the strongest argument for the role's existence is not a manifesto. It is a set of systems that are demonstrably doing the work.

Questions a marketing leader asks before creating the role

How is this different from a marketing operations manager?

Marketing operations runs the infrastructure. Marketing engineering builds new capability on top of it. An ops manager configures the CRM, manages the lead scoring model, and keeps the email platform running. A marketing engineer writes the integration that pulls ad performance data into the CRM automatically, builds the agent that drafts nurture sequences, and deploys the drift check that alerts the team when a pipeline goes quiet. Ops keeps the existing stack healthy. Engineering extends what the stack can do.

What happens when an agent breaks?

Every system that automates marketing work will eventually break. APIs change. Tokens expire. Platforms deprecate endpoints. The question is whether the break gets caught. A well-built marketing engineering operation includes a drift check: a scheduled process that compares what the documentation claims is running against what is actually running, and reports a defect count that should read zero. When the count is not zero, someone investigates before the gap becomes a campaign that did not launch or a report that did not send. The mechanism is the difference between automation that quietly rots and automation that stays alive.

Do we own what the marketing engineer builds?

A credible answer is yes, and it should be stated explicitly in the contract or employment agreement. Code, configurations, agent prompts, and documentation should live in the organisation's own repositories and accounts. The marketing engineer builds on the organisation's infrastructure. The organisation owns the output. Lock-in is a risk with any technical hire, and the defence is the same as it is with any engineering role: the work product is the company's property, and the systems are built in the company's environment.

How long before the first system ships?

A marketing engineer who is building from scratch, and who already knows the organisation's stack, can ship a working system within weeks. Best Medellin, the property portal mentioned above, went from a decision to a live site with four CMS collections, a homepage, every page template, and the first CMS-driven pages in a single day. That is an outlier, and it was built by someone who already had the scaffolding in place. A realistic expectation for a new hire: the first system ships within the first month, and it is small enough to be verifiable rather than ambitious enough to be impressive.

Is this a full-time role or can it start smaller?

It can start smaller. A fractional engagement that audits the existing stack and ships one high-leverage system is a low-risk way to test whether the role delivers before committing a full-time headcount. The audit itself has value: it maps what is manual, what should be automated, and what it would take, producing a written plan the organisation owns whether or not it hires further. The first system proves the concept. The full-time role scales it.

What does the role cost?

Rates vary by market, engagement shape, and whether the hire is full-time, fractional, or project-based. A full-time marketing engineer commands a salary at the intersection of senior marketing and mid-level engineering, because the role requires both skill sets. A fractional engagement is typically scoped as a systems audit, a build sprint, or an ongoing retainer. The scope is agreed before any work starts, so the organisation knows exactly what it is buying.

Join 200,000 marketers