Whoami?
Aka the quantum warlock, the masked camel, the fanOfThermodynamics, the pipe and clouds engineer, the whale orchestra conductor, or just Sckathach. I'm a french student at Télécom SudParis, and soon, a ML(4SEC) researcher!
I love mathematics, physics, graphs theory, AI, cybersecurity, graphs, geopolitics, thermodynamics (yes, I swear), programming, and especially graphs.
I'm currently fan of mechanistic interpretability, but there is a high risk I become fan of SLT instead.
My favorite paper (asof December 2024) : Toy Models of Superposition.
Education
May magically transform into a master of mathematics.
Telecommunications, cyber security, cloud, information theory, graph theory.
Experience
Mar - May 2025
Research internship in AI explanability - INRIA
Verified robust explanation for language models.
Jul - Dec 2024
Research internship in AI security - Thales
Implementations and improvements of state-of-the-art attacks on LLMs.
Creation of challenges (AI & quantum physics) and organisation of the 404CTF 2023 & 2024.
Papers
Using Mechanistic Interpretability to Craft Adversarial Attacks against Large Language Models - Winninger T., Addad B., Kapusta K. (mars 2025)
Cool logos to add colors to this page
I've put here a few logos of things I use, or get by with. Don't worry too much about suspicious ones.
