I'm Adrienne.

I build data platforms, pipelines and models, mostly for marketing and customer data. That often means connecting advertising platforms, web analytics, CRM and business systems. I also work out how their data should fit together.

I work on the business questions as well as the implementation. I define what a team needs to measure, integrate the sources and model the data, with maintenance in mind. I work mainly with BigQuery, dbt, SQL and Python.

From marketing to engineering

I studied marketing at Rotterdam School of Management and started my career in paid acquisition. I then worked on tracking and measurement, before moving into building the systems that collect and transform the data.

At L'Oréal, through Business & Decision, I worked on standardising Google Analytics and Google Tag Manager implementations across a programme covering 50 countries, 50 brands and thousands of websites. At that scale, consistent definitions and coordination were as important as the technical implementation.

At lePERMISLIBRE, I was responsible for marketing data within a five-person data team. My work included acquisition-cost modelling, comparing attribution models, and building real-time pipelines to route advertising leads to call centres. One of the questions we tackled was how much it cost to acquire a learner.

At UnNest, I work on data collection, ingestion, warehouse modelling and reporting for different clients. I adapt to each client's systems and to the people who will use and maintain them.

How I approach the work

My marketing background still shapes how I approach engineering. Before building a model, I want to understand what people will use it for and what its numbers are supposed to mean. A technically correct calculation can still answer the wrong question.

I also care about what happens after delivery. The next person working on a system should be able to understand where the data comes from, find the rules behind a metric and make changes without having to reconstruct the entire project.

That affects how I organise models, which tests I maintain, how I review changes and what I document. I make those decisions while building the system.

AI in my work

I use coding agents in my development work and build tools to support my data projects. I'm experimenting with ways to use them for source integrations, data transformations and documentation.

How much context does an agent need to work on an existing project? How do I test what it produces? Which actions can run automatically, and which need approval?

Why I write

I write about what I learn on projects and the decisions behind the code.

Some pieces explain an implementation in detail. Others start with a number, such as an acquisition cost, and examine what goes into it.

Read about the acquisition-cost project →

Outside work

I've volunteered with La SPA since 2019. I also enjoy learning languages, and I'm a beginner at embroidery.