How I know what I know.

Ten years spent getting interdisciplinary technical teams to agree on what to build next, in research labs and in companies.

I know what holds up, and how to find out.

Ten years of running experiments teaches you which claims to trust, which to test, and how to design the test so it settles something.

PhD in AI and Robotics, INRIA. National best-thesis prize, jury chaired by Fields Medallist Cédric Villani.
40+ peer-reviewed papers, 1,500+ citations, h-index 18. All listed on Google Scholar.
Published in Nature Communications, Science Advances, PNAS, Angewandte Chemie and NeurIPS.
Contributor to the French national AI strategy report.

Also: MSc in Micro and Nano Systems, highest honours; Diplôme d’Ingénieur (BSc and MSc) in Electrical and Computer Engineering, INSA Toulouse, best MSc thesis award.

I have owned this work, not only advised on it.

The gap is not ideas or execution, it is the handover between them. Those are usually different people, and I can speak to both.

Led a ten-person chemistry, robotics and machine learning team in a world-class lab: five postdocs, three PhD students, an engineer.
Set the research agenda on projects funded above $5M, and presented it to the stakeholders paying for it.
Tooling I built there is still in daily lab use a decade later.
Worked inside INRIA, the University of Glasgow, EPFL, UC San Diego, DTU and UT Austin.

I have tried to build companies, not only papers.

An academic who has never carried commercial risk gives different advice. I am not a serial founder with an exit behind me, but I have been on that side of the table.

Co-founded Pollen Robotics in 2016 around modular robotics. Patented the Robus modular bus and shipped a kit with publisher Nathan. I left in 2018, and came back as a consultant in 2024 until Hugging Face acquired them in 2025.
Founded SpyLock, turning my own research into a shipped Android product.
Entrepreneur First’s London cohort, and twice interviewed at Y Combinator.
Consulting since 2020, for clients from pre-seed to a $100M+ valuation.

I can hand the work over.

The arrangement I sell only works if your team owns the work after I stop turning up.

Trained the CTOs of growing start-ups through ML masterclasses: Morressier, R5, Tikweb, Remazing and Mursion.
Took Freetobook’s engineers from no machine learning at all to shipping it in production, growing from one developer in 2023 to three teams in 2026.
Sat alongside client engineers week to week, explaining the reasoning rather than handing down conclusions.
A decade teaching AI, data science and design thinking to cohorts of up to 400, and supervising 20+ year-long graduate projects.