Research Engineer
Safety, security and trust of AI agents: formal guarantees and adversarial red-teaming.
Research Engineer · AI safety & trustworthy agents
I'm a Research Engineer at Vijil, working on the safety, security and trust of AI agents, from formal guarantees to adversarial red-teaming.
Before that, I spent two years building open medical LLMs with the LiGHT lab (EPFL & Yale) and the MLO lab at EPFL, contributing to Meditron-70B and Llama-3-Meditron. I care about models that are useful where resources are scarce, which also took me to the International Committee of the Red Cross.
I hold an MSc in Data Science and a BSc in Computer Science from EPFL.


Research Engineer
Safety, security and trust of AI agents: formal guarantees and adversarial red-teaming.
Jun 2026 – Present
Research Associate
Worked on Apertus-Meditron-FO, a medical fine-tune of the Swiss LLM Apertus, and the Mediset dataset; training on A100 GPU clusters.
May 2024 – Nov 2025

MSc Thesis & Postgraduate Associate
LiGHT Lab, EPFL & Yale
Safety, instruction following and medical skills of Meditron LLMs (SFT + preference optimisation). Supervised 4 students.
Oct 2024 – Apr 2025

Intern
International Committee of the Red Cross
Multi-agent LLM chat for low-resource settings; ML for budget prediction and resource allocation.
Jul – Aug 2024

Co-Founder & Founding Engineer
Deal'icious (EPFL startup)
App for healthier, affordable food choices. Led vision, architecture and tech decisions.
Sep 2023 – Aug 2024
Research Engineer Student
MLO Lab, EPFL
Core contributor to Meditron-70B: instruction-tuning data, QLoRA fine-tuning and RLHF.
Sep 2023 – Jul 2024

Jul – Aug 2023
MSc in Data Science (GPA 5.5/6)
EPFL
2022 – 2024
BSc in Computer Science / Communication Systems
EPFL
2018 – 2022