AI × PRODUCT × SYSTEMS

I turn fuzzy AI ambitions into clear decisions.

Available for freelance work · Paris / Remote

How I work

From signal to system.

01Observe

Read the signal

Talk to users and prospects. Understand the market, the work and the constraints around the request.

IN / conversations, data, constraints
02Frame

Name the decision

Expose assumptions, narrow the opportunity and define what evidence would change our minds.

OUT / a problem worth testing
03Prototype

Make the loop tangible

Build the smallest credible slice that lets people react to the real idea, not a description of it.

RULE / tools follow the question
04Test

Put it in real work

Observe use and friction. Separate a compelling demo from something people would rely on.

PROOF / behaviour over opinions
05Integrate

Fit the system to the work

Connect context, tools and human validation. Document the loop so the team can actually run it.

OUT / a usable, owned workflow

Where we can start

Where do you need help?

A
Starting point

We have AI ambition, but no clear use case.

AI Opportunity & Prototype Sprint

Work
Research, map opportunities, prioritise, then prototype one narrow direction.
Outcome
A tested direction and evidence for the next decision.
B
Starting point

We want to improve an internal process with AI.

Internal AI Workflow Sprint

Work
Map the process, design the human and AI loop, connect context and tools, then test it with the people involved.
Outcome
A usable workflow, explicit human checkpoints and a grounded integration plan.
C
Starting point

We need someone to own the product direction.

Fractional AI Product PM

Work
Connect discovery, business priorities, technical constraints and delivery over time.
Outcome
An owned direction, clearer trade-offs and continuous learning.
D
Starting point

We have a product opportunity to frame and test.

Product Builder Sprint

Work
Clarify the angle, understand users and the market, then make the smallest credible version tangible.
Outcome
A testable direction and a practical build plan.

Selected evidence / 2020 to 2026

Things I've made real.

Some shipped, some stayed personal, one was deliberately stopped. Each left useful evidence behind.

HEEX TECHNOLOGIES · 2022 to 2026 · FIRST PRODUCT MANAGER

One role.
Three distinct layers.

I connected recurring client and user signals with business priorities and delivery constraints to reshape the roadmap and product direction.

01

Client Product

Continuous discovery, opportunity framing, prioritisation and delivery for the smart-data platform.

02

Product AI

Copilot experiments inside the customer-facing product, kept distinct from the core platform work.

03

Team Workflows

Custom MCP and assistant workflows connecting product knowledge, documentation and execution.

Heex analytics dashboard interface
Artefact 01 · product interface · dashboard capture

2020 to 2023 · Freelance client work

CLIENT WORKIN PRODUCTION

ADTE intranet

Role
Product builder
Problem
Field and administrative work was spread across after-sales, job-site, technician and office processes.
Made
A full-stack business intranet that remains in production, with later experiments around knowledge and operational assistance.
What it shows
Making software fit a real organisation, while staying responsible for the unglamorous details.

2024 to 2026 · Personal tools used in real work

SELF-INITIATEDINTERNAL WORKFLOW

Business AI experiments

Role
Designer and builder
Problem
Research, documentation and product knowledge repeatedly lost context between tools and people.
Made
Experiments around competitive analysis, documentation automation and internal assistants.
What it shows
Where automation can help a workflow, and where judgement must remain visible.

4 months of discovery · Self-initiated opportunity

STOPPED15+ INTERVIEWS

HR assistant

Role
Product discovery
Problem
An early hypothesis suggested a dedicated assistant could solve a recurring HR need.
Made
More than fifteen interviews. No launched product.
What it shows
General-purpose AI changed the need before launch. Stopping was the product decision.

Working file / Background

Where does the judgement come from?

Mines Saint-Étienne + emlyon

Engineering rigour + a business lens

Freelance full-stack developer

Learning to make the whole thing work

Veamly · San Francisco

A first product role, close to users

First Product Manager · Heex

Discovery, direction and delivery under real constraints

Freelance Product × AI × Builder

Turning unclear decisions into testable systems

Next decision / Contact

Have a product or AI
challenge in mind?

Share a little context with me on LinkedIn or by email. I'll be happy to take a look and tell you honestly whether I can help.