Hire — Yoann Leny

Target roles: VP Data · Head of Data & AI · Chief Data Officer
Base & mobility: Bordeaux, FR · remote-first · regular travel across Europe & North America
Languages: English · French
Availability: Full-time start within a normal notice period

A strong fit when

The data function exists but doesn't compound — tools everywhere, trust nowhere.
Agentic AI is on the board agenda and someone has to make it real, governable, and priced.
The team is 10–60 people and the operating model, not headcount, is the constraint.
You want an operator who ships and runs systems, not a strategist who hands over a deck.

A poor fit when

The role is a caretaker position — keep the dashboards on, change nothing.
AI is wanted as theater: demos for the board, no appetite for governance or evaluation.
The mandate has responsibility for outcomes but no authority over the operating model.

THE CONCRETE PART

My first 90 days, in writing.

Every candidate says "first I'd listen." Here is the actual plan I run, phase by phase — the same one I have used to take over data functions before. Click a phase to expand it.

Days 1–30Listen, map, and find the constraint

One-to-ones with every direct report, key stakeholders, and the loudest internal critics of the data function.
Map the real system: data flows, decision flows, and where they diverge from the org chart.
Audit the estate — platforms, pipelines, metric definitions, AI initiatives — for trust, cost, and ownership gaps.
Identify the single operational constraint that, if removed, changes the trajectory. There is always exactly one that matters most.

DeliverableA written diagnostic the executive team can argue with: what works, what doesn't, what it costs, and the one constraint I propose we attack first.

Days 31–60Ship one visible win, design the target system

Deliver one fix executives can see — usually a trust win: two dashboards that finally agree, one metric with a contract, one AI workflow with guardrails.
Design the target operating system: platform architecture, semantic governance, team topology, and the agentic roadmap — as one design, not four documents.
Set the metric baseline: decision latency, data trust incidents, cost per outcome — so progress is measurable rather than narrated.
Make the first hard ownership calls: every critical data product gets a name next to it.

DeliverableOne shipped improvement, plus a target architecture and operating model the team helped shape and the board can fund.

Days 61–90Install the operating rhythm

Stand up the operating cadence: monthly operating review on real delivery-health data, quarterly topology review, metric change process.
Start the first structural build from the target design — with the existing team doing the building, not watching consultants do it.
Publish the 12-month roadmap with explicit bets, costs, and the assumptions that would change them.
Report back against the day-30 diagnostic: what I said, what I did, what I got wrong.

DeliverableA running operating system — cadence, ownership, roadmap — that no longer depends on me being in the room for every decision.

EVIDENCE, NOT ADJECTIVES

Where to check my claims.

Case studies

Four production engagements with outcomes, architecture decisions, and what I would do differently.

Read the work →Decision log

Dated technical bets I made and how they aged — including the ones that didn't.

Check my judgment →Concept library

50+ maintained entries on data architecture, engineering, BI, and agentic AI — with named opinions.

Test my depth →

If the mandate is real, let's talk.

Email me about a role →Reach out on LinkedIn

Full CV available on request — the one-page version arrives within a day.