Current position

I am a PostDoc within the Cronin group. The group study the origins of life from different perspectives. I am particularly interested by the intersection of chemistry, robotics, and AI.

I am leading the Chemobots team that comprise 2 PostDoc, 4 PhD students, and 1 intern. Our objective is to combine, develop, and evaluate algorithms for exploration of complex physicochemical systems. Such systems cannot be simulated in practical times and experiments must thus be performed on the real system – raising a number of interesting challenges. We take inspiration from the field of developmental robotics, with concepts such as goal babbling, intrinsic motivation, and maturational constraint, and apply them to the exploration of complex systems in the real world.

We recently demonstrated the use of robotics and artificial intelligence as tools to explore oils in water systems. A robot equipped with an evolutionary algorithm was able to optimize efficiently oils ratio to elicit various droplet physical behaviors, such as movement, division, and vibration.

We are now exploring the open-ended exploration of such systems and the role of the environment as an experimental variable impacting the expression of physicochemical systems.


I defended my thesis titled “Learning from Unlabeled Interaction Frames” on October 24th 2014. I pursued my PhD at the INRIA and Ensta-ParisTech Flowers Team where I investigated how to create calibration free interactive systems. I was advised by Manuel Lopes and Pierre-Yves Oudeyer. I received the ''Prix Le Monde de la Recherche Universitaire'' 2015 for my thesis work.

I am also a maker and an active member of the Poppy project, an open-source project providing tools to enable the creative exploration of interactive robots for science, education, and art. In 2015, we founded an association called "Des Robots En Baskets" to promote robotics as a tool for transdisciplinary education at many levels.


I received the ''Prix Le Monde de la Recherche Universitaire'' 2015 for my thesis work. Each year, this price is awarded to 5 young French scientists across all fields of science.

[More inforation]

I am highly involved in Poppy, an open-source project for the creation, use and sharing of interactive 3D printed robots. I actively contribute to the expansion of the project, participating in board meetings and developing technological proof-of-concepts.

Parametrizable robots

I have been developing open-source libraries for creating parametrizable robots: [lego-scad], [raspberry-scad], [robotis-scad], and [segment-scad]

That has been use for the creation of the following robots:

The following robots also started with openscad design:

The latest version are now using an onshape design by Matthieu Lapeyre. See the hardware/ file of each repo to find back the openscad version.

See the project page

Online visualisation

I started developing a web visualisation for the ErgoJr robot to allow the community to discover our tools without access to a physical robot. The visualisation has since been greatly extended and integrated within the Poppy ecosystem by Théo Segonds and Damien Caselli.

It is available online at

Soft robot

I developed a soft connector for XL320 based robot, making it easy and cheap to design and include soft elements in Poppy creatures. I designed a Starfish robot to showcase the use of this connector. More information on the Poppy forum:

See the project page

Our PlosOne paper on self-calibrated brain-machine interface control using error-related potentials is online. The code and the paper source are available online:

This work is a direct application of my thesis work on learning from unlabeled instructions. See the project page and my thesis defense.

[Main Publication]