I'm an AI Research Engineer at Helsing, a new type of defence company delivering AI-enabled autonomous systems to protect democracies. My work focuses on development of safeguard models and scalable evaluation of our AI products, ensuring that AI components operates reliably in real-world environments. I also co-founded Future Frame in 2024, a startup building foundation models for tabular data, backed by Entrepreneur First and Transpose.

I hold a PhD in Machine Learning from CentraleSupélec at Université Paris-Saclay, where I published research into making AI safer to operate in real-life applications, focusing on out-of-distribution detection, uncertainty quantification, and selective prediction, supervised by Pablo Piantanida and funded by BPI France and IBM Research. I did a research internship at Mila in Montreal in 2022, working on uncertainty quantification for medical image segmentation. I hold an MSc in Applied Mathematics and Data Science from ISAE-SUPAERO in Toulouse, and a BE in Aerospace Engineering from the Instituto Tecnológico de Aeronáutica in Brazil.

As a side project, I co-created Prompt Riddle in 2018, an online enigma-solving game with over 5,000 players, where the goal is to escape an emulated terminal by outsmarting an AI.

Broadly, I am interested in developing robust general purpose AI models.