Senior Marketing Technologist

If you take this role, here's what's true 12 months from now. Our CDP is live on Databricks and activating audiences to multiple channels. The identity-resolution model you built handles deterministic and probabilistic stitching across our ecosystem and brand portfolio. AI agents draft audiences, briefs, copy variants, and stakeholder reports — and you've measurably raised the quality of their output by owning the human‑in‑the‑loop QA. When a user triggers an event, that signal reaches every activation channel within four hours. The global and talent marketing teams across the regions activate new campaigns by themselves. Direct sourcing audiences are live and we can measure bid‑to‑placement uplift by segment.

The spike we're hiring for: We don't need a generalist. We need a senior operator unusually good at one thing: orchestrating identity, audiences, and activation on a Databricks‑native data platform — writing the PySpark and SQL that makes it work — while raising the bar on an AI agent layer that drafts work for you.

Why this role exists: HeadFirst Group is building a new function from scratch, inside the Product organisation. The objective is to leverage our data fast, to enable product‑led growth across our marketplaces and brand portfolio, starting with direct sourcing / direct fulfilment activation, global campaigns by job category, and personalisation at scale. This is not a marketing campaign role. The global marketing team and the talent marketing teams across the regions own content, channels, and campaign execution. You provide the data foundation, activation tooling, and experimentation capabilities that let those teams scale performance consistently. You are the first hire into the team. Your manager runs strategy, architecture, vendor selection, stakeholder politics, and the agent roadmap. You own the activation layer — config, identity, events, and the day‑to‑day quality of what ships out of the agents.

How your time splits: CDP configuration and audience design — 50%. Identity resolution and event taxonomy — 25%. Event tracking across our platforms — 15%. Agent QA and prompt iteration — 10%.

The agents you'll work with: By the time you start, three agents are already live in production. Five more come online in your first six months. They draft audiences, briefs, copy variants, segmentation logic, prioritisation calls, and stakeholder reports — the kind of work that used to consume a team. Your job is the 20% of judgement those agents can't do: identity calls, consent decisions, and the call on whether an agent's output is good enough to ship. You are the human in the loop on every QA cycle. If you read 'AI agents draft your work' and feel relief, this isn't your role.

What 'good' looks like to us: We hold ourselves to the bar of teams that have publicly figured out warehouse‑native MarTech: composable stacks, the GTM engineering shift, operations as orchestration. We notice when a marketing engagement tool list quietly drifts away from the rules we defined for it. We argue about identity stitching. We think a Job Description should be a stake in the ground, not a checklist.

What you bring/What you can show us: A reverse‑ETL flow you shipped to production. Not training; production. A dbt model you wrote and got merged into production. Not a course you took — code that shipped. A PySpark job you shipped to Databricks production — pipeline code, not a notebook demo. An identity‑resolution decision you made — deterministic vs probabilistic, with the trade‑off you accepted and what broke later. One example of correcting AI‑drafted output that would have shipped wrong, with what your judgement caught.

The stack you'll work in: Data platform: Databricks CDP / activation: Hightouch / GrowthLoop ESP: HubSpot for B2B, Braze / Iterable under evaluation for Business to Talent. CMS: Sanity headless.

What we offer: €6,500–€8,500 gross monthly base salary (Schaal 7), senior profile, plus performance bonus. Job Family: Product. Full‑time contract — one‑year fixed term first, with the intent to convert to a permanent contract. 27 vacation days with the option to buy more. 8.33% holiday allowance and premium‑free pension. Hybrid working with flexible hours and home‑office allowance. Modern office in Hoofddorp. Access to HeadFirst Academy and internal learning programs. AI tooling funded by default — Claude, Cursor, whatever you need. Vendor budget already committed — you build, you don't sell internally.

How you operate: You read SQL fluently and write the JOIN that debugs an audience definition without asking. You write Python and PySpark on Databricks as core craft — transforming bronze to gold, building identity logic, productionising notebooks. Not a plus. A requirement. You talk in audiences, traits, and events. Not lists, properties, and campaigns. You can sketch a server‑side consent flow on a whiteboard at high level, without notes. You're comfortable being the sole owner of tools nobody else in the company yet understands. You say no to one‑off regional requests that don't fit the audience model — and you can make that no, land politely.

Helps but not required: Reverse‑ETL plus warehouse‑native CDP combination in production. Multi‑brand or multi‑market activation experience. Headless CMS experience in production. Experience working alongside AI agents in production — not just prompting Claude. Dutch — most of the team operates in Dutch. Nice to have, not a requirement.

Not what we're looking for: A HubSpot or Marketo admin who wants to expand into MarTech. The mental model doesn't transfer. A 'marketing automation specialist' whose deepest tool is a marketing‑cloud UI. Someone who treats AI agents as competition, or as a magic 10× button. Both lead to the same bad outcome. A vendor‑demo enthusiast — the major architecture decisions are already made. A campaign manager — that's a different role. Anyone who needs a 12‑vendor RFI before they can make a call. Anyone who needs to be told what good looks like.