join us

Rhizome Labs

From the blog →

About Rhizome Labs

Our purpose is to build trustworthy adaptive intelligence for demanding real-world environments: AI systems that can run locally, protect sensitive data, remain reliable as conditions change, and support responsible human oversight.
We are developing task-specific models designed for deployment beyond the lab, where resilience, transparency, and operational confidence matter as much as raw performance.
As passionate advocates for open-source, we aim to contribute to this new path forward by promoting open and rigorous science for safe, useful, and accountable AI.

Science without conscience is but the ruin of the soul

Rabelais


Who we are
We are Lancelot (CEO) and Clement (CTO), two repeat founders bringing together product execution, company building, and deep technical research. Lancelot leads strategical partnerships, and product development. Clement leads research and engineering, with a Ph.D. in Neurosciences and Deep Learning from CEA/Inserm and experience at the intersection of biological and artificial intelligence.

How we work
We believe in radical transparency, boundless ambition, and relentless curiosity. Rhizome Labs is defined by the intellectual richness of its members; a flat organization where every voice matters, and where open dialogue and mutual respect cultivate a healthy, idea-driven environment.
We are hiring research engineers. Join us!

Partners
Feel free to reach out at partnerships@rhizome-labs.com if you are interested in collaborating on trustworthy adaptive AI systems for privacy-sensitive, industrial, or operationally demanding use cases.

Publications

  1. Yassir Bendou, Omar Ezzahir, Eduardo Fernandes Montesuma, Gabriel Mahuas, Victoria Shevchenko, and Mike Gartrell. ReBaPL: Repulsive Bayesian Prompt Learning. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026.

    arXiv:2511.17339
  2. Mariana Vargas Vieyra. Staying Alive: Uncensored Survival Analysis with Tabular Foundation Models. arXiv preprint, 2026.

    arXiv:2606.03689