About me.

My working philosophy is simple: clean data isn’t an end in itself. It’s what lets marketing teams argue with evidence, stop mistrusting their dashboards, and operate on the decisions they make. The data work only matters if someone, somewhere, makes a better call because of it.

I work with AI, not despite it. Claude Code and custom agents are baked into how I deliver — which means I ship faster, with less overhead, at the same quality. Before this was a thing people talked about, I was writing my own dbt-generating skills and internal tooling. That’s a real edge, not a marketing line.

How I work

• Specificity over generality I don’t take on projects I can’t execute well. If what you need isn’t in my zone, I’ll tell you and point you somewhere useful.
• Fixed scope, fixed price where possible Consulting engagements that run on day rates alone tend to drift. I price audits and tracking setups flat, so you know what you’re getting.
• Teams first, tools second Stacks should be handed off, not inherited. Documentation and training are part of every engagement.
• Honest about AI AI changes what I deliver. It doesn’t change what matters. I won’t sell you an ‘AI transformation’ because that’s not a real thing — but I will build you workflows that save your team real hours.

Background

Before going independent, I spent several years at unnest, a Paris-based data consulting agency, working with brands including ba&sh, Nutrimuscle, PrivateSportShop, Lafuma Mobilier, and PSG. I’ve handled everything from dbt migrations to server-side tracking rebuilds to CNIL-compliant consent implementations.

I keep working notes on the modern data stack, dbt, and the overlap between data engineering and AI. Read some notes.

Outside work

I volunteer with La SPA, a French animal welfare organization, where I donate my data and marketing skills. I’m also learning several languages (Russian, Japanese, Korean, Chinese — with varying success).