I've been coding since I was a kid, mostly because I liked taking tech apart and figuring out how it worked. The fascination with machine learning came later.
My first engineering experience was at Pernod Ricard, where I learned what it actually takes to put a project into production. Working with messy data at scale, adapting to business stakeholders with their own constraints, caring about data quality in ways nobody really teaches you. It changed how I think about ML work. These days I also freelance on CookiNUM, a low-latency streaming platform, now running across 20+ institutions.
Research has always pulled at me too, which is why I'm spending six months at the RALI lab at Université de Montréal, mapping NLP benchmarks across 74k+ ACL papers. It's been clarifying. I love reading the papers, but what I actually want to do is turn that work into production-grade ML systems people can actually use.