About me

This page is an informal introduction to me and what I’m up to.

A more formal introduction is at CV.

I try my best to maintain a page listing the tools I use during my psychiatry residency: TOOLS.

Open for collaborations. If anything on this page resonates, please get in touch. I’m open to all sorts of collaborations: paid or pro bono, small one-off questions or full-blown research projects, and everything in between. My “expertise” is largely self-taught, but I’d rather hear about your idea and tell you honestly than have you not ask.

Context

Since I was about 10, I’ve wanted to do psy-something. A few years later, that crystallized: psychiatrist was my ultimate dream job.

But I couldn’t give up computer science. So, considering the quirks of the French education system, I decided to combine both my way.

Education

I took advanced math/physics/chemistry classes in classe prépa, then went to medical school during which I did a 2 months master research internship (French: “stage de M1”) at NeuroSpin creating from scratch a cluster plotting and benchmarking library for neuroscience. In 2025 I started medical residency in psychiatry in Paris, France. The same year, I created oliCorp, a one-person company (French: “auto-entreprise”). This was necessary for legal reasons when providing services to a French non-profit.

I started coding when I was about 12. I’ve always been deeply committed about open source - every time possible I make my code publicly available on my github. Here are my stats:

I have experience working in collaboration and autonomously: I had the opportunity to build innovative solutions from scratch and make them freely available. Examples of my self-motivation include:

• Tinkering with high dimensional embeddings led me to create QuestEA and pitch it to renowned researchers. The goal being to uncover hidden patterns in medical questionnaires allowing to unify partially disjointed medical datasets.
• To query my large collection of medical documents using natural language, I created wdoc, a RAG-based tool that can ingest dozens of file formats and recursively summarize arbitrarily long documents. It proved invaluable for quickly finding official sources in high-stakes medical situations, and has since been used by lawyers and other professionals.
• Tinkering with AI interpretability led me to fork the repeng library to test various high dimensional vector modification techniques to steer its behavior.
• Getting frustrated with my sleep, I created from scratch SleepTk, a sleep tracker app for an open source smartwatch with an innovative pattern-based vibration alarm to wake me up gently and without waking up my partner. After years of use I organized a bug bounty to rewrite it in C on an optimized OS to get a much better battery life. Still using it every night.
• More at PROJECTS.

Interests

My core interest has always been around neuroscience with a focus on psychiatry.

Circa 2015, I got interested in AI/ML so naturally I’m heavily drawn towards computational psychiatry.

I’m excited about how AI research will advance psychiatry research, improving patient’s lives for the better.

At the same time, clinical observations from psychiatry could inform more robust and interpretable AI systems that better serve human needs.

Things I love (besides caring for my patients)

In no particular order:

• Frugal innovation. The idea that small, agile, multitasking people can occasionally build useful things with limited means.
• Medicine. Both to help others and to know thyself. And because being with my patients is something I absolutely love!
• Open source software/hardware (including this very website). I am part of InterHop (and donator as well!).
• When building blocks I’ve participated in have positive compounding effects.
• Applying the Golden Rule.
• Bayesian thinking (see Julia Galef’s explainer video).
• Repairing stuff.
• Robust, distributed, decentralized systems.
  • More

    • I have a deep fascination for emergence and self-organization
    • The former essential, I think, to understand counsciousness. And the latter is very connected to my interest in distributed and decentralized systems (e.g. AT Protocol, Reticulum, PRIME-IROH, edgevpn, Nostr, Conflict-Free-Replicated-Data-Type, ...).
    • In my view, both are deeply linked to neuroscience and will be useful to help patients.
• Encryption-by-default and privacy respecting tech.
• Self hosting services for me and people around me (including this very website).
• HackerNews.
• The IndieWeb movement.

Other

Bug Bounties Payment Track Records

I actively support open source development through bounty incentives and community contributions:

1. Porting my SleepTk app from a micropython-based OS to a C-based device.

2. Adding highlights on the mobile webapp for Karakeep.

3. Additional bounties are underway.