I am a PhD engineer working at the intersection of AI, optimisation and complex industrial systems. During a four-year industrial PhD at CentraleSupélec and Safran Electronics & Defense, I developed methods combining neural Lyapunov functions with global optimisation to identify stability regions in nonlinear system models. I then turned this research into a tool Safran engineers could use without writing code. The work led to three peer-reviewed international conference papers as first author.
More recently, I co-founded Hoocq (opens in a new tab) and helped grow it to €300k+ in revenue across 50+ clients, combining business development, client strategy and end-to-end project delivery through a network of around 100 freelancers.
I am now looking for technically demanding AI problems where scientific rigour and real-world execution matter equally.
Selected projects
python · pytorch · neural certificates · ga / pso
A Python and PyTorch reimplementation of the methods from my doctorate for automatically searching Lyapunov functions and estimating stability regions in nonlinear dynamical systems. Neural Lyapunov functions and quadratic forms are optimised with genetic algorithms and particle swarm optimisation, for both continuous- and discrete-time systems. Reproducible experiments, tests and CI.
Case study
·
Code (opens in a new tab)
Experience
since 2023
Hoocq
Co-founder. Led business development, client strategy, budgeting and end-to-end project delivery, coordinating projects through a network of around 100 freelancers. Integrated generative AI and workflow automation into prospecting, content production and selected client projects.
2020–2024
Safran Electronics & Defense × CentraleSupélec
Industrial PhD researcher (CIFRE). Developed optimisation-based methods to construct neural Lyapunov functions and identify stability regions in nonlinear dynamical systems. Combined genetic algorithms, particle swarm optimisation and genetic programming, validating the methods on Safran industrial models with up to nine state variables.
2021–2023
CentraleSupélec
Teaching assistant in control theory. Taught control theory and MATLAB practical sessions to second-year engineering students.
2018
Oslo Metropolitan University
Research engineer. Developed Python behaviours for Thymio II robots in a swarm robotics experiment, combining random exploration with visual trail-following to mimic pheromone-based coordination.
Publications
A comprehensive framework to determine Lyapunov functions for a set of continuous time stability problems
B. Bocquillon, P. Feyel, G. Sandou, P. Rodriguez-Ayerbe
IECON 2022 — 48th annual conference of the IEEE Industrial Electronics Society · Brussels
IEEE Xplore ↗ · HAL, open access ↗
Computation of neural networks Lyapunov functions for discrete and continuous time systems with domain of attraction maximization
B. Bocquillon, P. Feyel, G. Sandou, P. Rodriguez-Ayerbe
IJCCI 2020 (NCTA) — 12th international joint conference on computational intelligence · pp. 471–478
doi.org/10.5220/0010176504710478 ↗
Efficient construction of neural networks Lyapunov functions with domain of attraction maximization
B. Bocquillon, P. Feyel, G. Sandou, P. Rodriguez-Ayerbe
ICINCO 2020 — 17th international conference on informatics in control, automation and robotics · pp. 174–180
doi.org/10.5220/0009883401740180 ↗
Training methods for analyzing the stability of a complex system
PhD thesis — CentraleSupélec, Université Paris-Saclay, with Safran Electronics & Defense
Defended 25 March 2024
HAL, open access ↗
Education
2024
PhD in AI, Optimisation & Control
CentraleSupélec · Université Paris-Saclay
Industrial PhD conducted with Safran Electronics & Defense
2018
Engineering Degree
ISEP, Paris
Specialisation in Business Intelligence
Academic exchange at Oslo Metropolitan University, Norway