Projects

I believe that navigating the transition to a world with strong AI is the most critical challenge of our time. However, despite the rapid advancement of AI capabilities, the field of AI safety remains significantly under-resourced. I founded Factor Omega to help address this imbalance and push the field forward.

Factor Omega is a non-profit organization dedicated to accelerating projects that mitigate AI risks. By reallocating strategic resources like GPUs, we aim to bridge the gap between AI safety research and practical applications.

Our mission addresses the significant disparity between the rapid advancement of AI capabilities and the comparatively limited investment in AI safety. We seek to redirect underutilized resources into early-stage, underfunded AI risk reduction areas to ensure a safer AI future.

I am proud to be a founding member of Worldcoin, a trailblazing cryptocurrency project with a vision of establishing a more equitable distribution of wealth across the globe, as well as laying the foundation for a novel identity primitive in the imminent AGI-enabled internet era.

Our initial team, composed of seven plus Sam Altman (Chris Brendel, Alex Blania, Fabian Bodensteiner, Max Novendstern, Philipp Sippl and Sandro Herbig), started this project in a very Silicon Valley way, setting up camp altogether in a small apartment and working 24/7 together.

Along the way, we have had to work on numerous and diversified issues. Building a state of the art biometric imaging device (the orb), a new cryptocurrency, a new decentralized identity system, and being privacy preserving throughout all of our tech.

An integral part of the success comes from the resilience, dedication and intensity of the team we have put together. We are now well on our way towards our goal; signing up over a billion people to Worldcoin.

A few years after my London adventure, I made my way to Caltech to delve into AI and Neurotech, though I also dabbled in Math and Physics just to keep things spicy. Caltech easily had the highest concentration of mind-boggling talent I had ever encountered—people who seemed able to juggle PDEs, CRISPR breakthroughs, and cosmic inflation theories before lunch.

The bar was set ridiculously high; even the least intimidating person I met could still recite half of quantum mechanics from memory. Everyone was a polymath in some shape or form, which meant you'd start a chat about neural networks and somehow veer into the ethics of lab-grown organs, then loop back to black hole thermodynamics. It was equal parts thrilling and surreal.

My time at Caltech overlapped with the COVID era, turning campus into a weirdly perfect playground for uninterrupted learning. Best of all, it was where I met the friends and collaborators who would later become the core team behind Worldcoin; so I guess that social distancing paid off.

My tech journey officially kicked off at Imperial College London, where I dived headfirst into Computer Science and AI. When I arrived in 2014, my English was, shall we say, "très mauvais", and my French accent made every sentence sound slightly comedic. Still, I gradually found my footing in the bustling metropolis of London.

Imperial was a melting pot of brilliant minds and infinite late-night lab sessions. It was here that I learned computers aren't just fancy calculators but powerful tools that could shape entire industries—and maybe the world if you're daring enough.

Outside the lectures, I also discovered the joys of university parties, forging lifelong friendships among neon lights and pounding music.

Simultaneously, I tackled a remote Bachelor's in Economics and Management back in France because apparently I thought my schedule was not busy enough.

During those years, I also tutored high school and university students on the side to help keep me afloat financially. That juggling act of labs, tutoring and extra classes was intense, but it molded my work ethic in a big way.

I've always been passionate about science and history, and I've long thought the most engaging way to learn history would be through conversations with historical figures themselves. Now that technology has caught up, it felt like the right time to make this idea a reality.

I built a system that generates podcast episodes featuring immersive and in-depth discussions between a curious young time traveler named Leo and some of history's most notable figures.

Initially, I created these podcasts just for myself and started sharing them with family. But they turned out pretty enjoyable, so I decided to share them with the world. I also recently open-sourced the entire automation pipeline on GitHub for anyone interested in how it works.

Back in 2023, I felt that chatbot interfaces like ChatGPT and Claude, while powerful, lacked deeper context, proactive behavior, and the ability to actually take actions on your behalf. So I built a custom AI friend for myself, MirrorAI.

At its core, MirrorAI drew on large language models like ChatGPT, but I bolstered it with integrations across Notion, Google Workspace, AWS, GitHub, messaging apps, emails and more. Privacy was built in too; no backend, it talked directly to the APIs.

MirrorAI is now happily retired. Claude Code came along and achieved exactly what I had been trying to build, only better (kudos to the Claude team).

After years of being on the founder side of the table, I figured it was time to see what it's like from the other side. Frst is a seed-stage fund with roots in Paris and a growing presence in SF, so it felt like a natural fit.

Mostly, I just enjoy meeting ambitious people building weird and wonderful things, and occasionally being useful to them. Turns out that's basically the job description.

EWOR named me one of their Top 100 Founder Mentors, which was flattering. Then they asked if I wanted to help spot exceptional founders as a Venture Scout, which was even more flattering. They run one of the most selective founder fellowships out there (35 accepted from 35,000 applicants), and are now expanding into the US, which is where I come in.

The interesting part is that EWOR backs people before they even have a company. Just a dedicated mind and something worth chasing.

My job is to find those people, which is basically what I was already doing, except now someone keeps score.

I have been experimenting with what might be the laziest form of entrepreneurship: businesses that are conceived, built, and operated entirely by my fleet of AI agents. The idea is to test whether autonomous agents can handle the full lifecycle of a micro-business, from product development to marketing to customer support, without me lifting a finger (or at least, fewer fingers).

The way I see it, in 2025 we are about a year away from anyone being able to spin up a real business with nothing but a laptop and a few well-prompted agents. So naturally, I had to try.

Current experiments include VisionTags, a Shopify app that auto-generates product tags and SEO metadata from images using AI vision; GEO Compare, which blind-tests how AI models perceive and recommend your website versus competitors; and a handful of web scraping actors on Apify.

In the summer of 2018, I decided to do something that sounded less insane in my head than it ended up being in real life: walk across Morocco for nearly a month. Starting near Essaouira and finishing at Fès Airport, my brother and I covered roughly 750 km the slow, sweaty way—on foot.

We had the usual suspects working against us: searing heat, blisters that sprouted their own blisters, and enough local dogs to keep us constantly scanning for fangs. But the unexpected ace up our sleeves was Moroccan hospitality.

By the time we limped into Fès, we were exhausted, dusty, and absolutely certain that humans are mostly awesome.

In the summer of 2016, I decided my legs were underused and embarked on a two-month, 5,500-kilometer pedal-fest from the fjords of Bergen, Norway, down to Rome, Italy; mainly because I found it hilarious to say "all roads lead to Rome" like a mischievous fourth grader.

Our team consisted of me, my 16-year-old brother Auguste, and my cousin Clement. We cycled through Norway, Sweden, Denmark, Germany, Czech Republic, Poland, Austria, Slovakia, Slovenia, Croatia, and finally Italy; knocking out about 97 km a day.

In the end, it was equal parts ridiculous and enlightening; and now I can smugly say I actually biked to Rome.