About Me

Hey, I’m Ilyass Ouardi!

I’ve always been drawn to challenges. As a kid, math was my favorite subject, not despite being hard, but because of it. Where others saw something to avoid, I saw a puzzle waiting to be solved. I’d watch movies about geniuses cracking impossible problems and then go look up unsolved math questions, half-dreaming I’d solve one myself. Naive, sure, but that instinct, to chase the hardest open question in the room, is exactly what pulls me toward research today.

Computer science became the natural playground for that curiosity, and over the last year it has sharpened into a clear direction: I want to understand what actually happens inside large language models. We’ve built systems that do extraordinary things while barely understanding how they work, and I’ve decided to spend my energy on closing that gap — the field of mechanistic interpretability, and the broader question of what makes these models reliable, transparent, and safe. What I love is that this sits right where computer science turns razor-thin against math: elegant, formal problems whose solutions read almost like mathematical poetry.

For me research isn’t just an aspiration I talk about, but something I’m already experimenting with. In the last academic year I’ve carried three research projects from question to result as projects for my exams, the most important of which studies the internal geometry of “truth” in LLMs and is now becoming my Master’s thesis, in collaboration with a PhD student and a professor at my department. That work, and the others, led three different labs to invite me to collaborate. My next goals are simple and ambitious: a strong thesis, a first publication why not at a top venue, and a PhD in mechanistic interpretability.

Outside research, my other great passion is football, I’ve played since I was a kid, from the streets to semi-pro, and I live for the competition. Basketball is how I switch off after a long day of studying, and I’ve even done some modeling, which introduced me to great people and a different take on creativity. I think the same thing drives all of it: the thrill of pushing myself and getting better.

If any of this resonates, mech interp, AI, hard problems, or just the drive to take on something difficult, reach out. I’m always up for a good conversation, a new project, or trading ideas. <!–

Hello, my name is Ilyass Ouardi.

I’m a Computer Science student at the University of Milan, driven by a passion for blending technology, math, and a touch of social good. My journey into the world of problem-solving began with high school coding and math competitions, and that spark has only grown since. From developing tools to assist non-profits to diving into the depths of machine learning, I’m always on the lookout for ways to push the boundaries of what tech can do.

Research is where my heart lies, and I’m particularly fascinated by machine learning, theoretical computer science, and software engineering. My goal is to contribute to cutting-edge advancements and pursue a PhD in the United States.

When I’m not buried in code or research papers, I’m on the field—whether it’s playing football or shooting hoops. I’ve also dabbled in modeling in Milan, where I’ve met some amazing people and broadened my perspective.

Thanks for stopping by!

–>