How I got here.

Before

Growing up, the thing I was best at was school, and the thing I was worst at was sitting still in it. I was first in my class, first in my whole grade, and most of my teachers couldn't stand me. I talked constantly, walked around during lessons, and was generally annoying. In retrospect, they had a point.

Around fourteen I decided decent was good enough, and it was, until grade 11. Then I locked in for a year and a half, just long enough to get into the university I wanted, and not a day longer.

What I'd study there was a different question. I never knew what I wanted, and I hated studying, which I still do. So I did the French bac in grade 11, started grade 12 in the math section, and moved to biology a week or two later. Then I applied to seven majors at four universities. I got into mechanical engineering at AUB, and two semesters later I switched to electrical and computer engineering. If there was a plan, nobody told me.

2019–2021

By then, the comfortable life I'd grown up with had been draining away for years, and the 2019 crisis finished it off. I was twenty and in my second year at AUB.

Early in 2020 I started learning machine learning. I'd like to say I had some grand vision, but mostly it just seemed like something I should look into. So I did the standard courses: Andrew Ng's Machine Learning, then his Deep Learning Specialization.

What actually got me was Magenta, a project at Google that was trying to get neural networks to write music. The music was terrible, and anyone could hear that. But I'd been playing music my whole life, and a computer writing any at all was mind-blowing.

After that I didn't need a reason. I did the GAN courses, read everything I could find on VAEs, and looked at more mel spectrograms than I can remember. That year I also joined an ML lab at AUB that was working on Arabic BERT models.

In October and November 2020, I lost both of my grandmothers, and I left the lab soon after. In February 2021, I lost my father.

I wasn't in a great place even before all of this, and for a while it broke me. Coming back took a long time, and it happened slowly. I've supported my family, emotionally and financially, the whole way.

2022–2023

In 2022, through my mom's network of all places, I was introduced to a Stanford grad who was starting a company. He asked me to write something about who I was and what drove me. This was before ChatGPT, so I actually had to write it. He liked it, but there wasn't a fit yet.

A year later I was a few months from graduating and had no idea what came next. I only knew what I didn't want: a master's, or a job at one of the big Lebanese firms. Then one night at 11pm, while I was in a lab at AUB, he called. They were pivoting and needed an AI engineer.

I did a trial, met his cofounder, and joined Gaya before it had any revenue. It has since made millions and turned a profit.

2023–2025

Alongside Gaya, I helped build a handful of other startups, as a founding engineer, a contractor or an advisor. Between them they've raised and made millions, backed by top investors in the US and MENA.

But every one of them was someone else's.

A job is a trade. You give up most of the upside, and in return, when things go wrong, it isn't on you. For most people it's a good trade. It never was for me. I'd rather own the whole thing, including the chance that it crashes and burns.

I value freedom the same way, though not in the sense people usually mean. I work a lot. But a lot of work on something you chose is a completely different thing from a little work on something you didn't.

So I left Gaya to build my own thing, and swore I'd never be "employed" again.

2026

In early 2026 I started Clavis. We were a small, lean team doing work for the kind of very large firms that agencies usually only dream of. It was working, and I had no plans to leave.

Then, one day in Tbilisi, the other Joe called. I'd known him and Neahme, now my cofounders at Bilo, since AUB's iPark, and I'd always liked what they were building. But the problem had turned out to be much bigger than they'd thought: the layer every out-of-home campaign in the world runs through. A good business doesn't hold up next to that.

In June I joined them as CTO and cofounder, and spent the summer winding Clavis down. It was the ownership I'd been after all along. And technically, I kept my promise.