About Libertalia
Named after the legendary pirate utopia, Libertalia is an independent research collective built on principles of freedom, openness, and collective endeavor.
“No gods, no masters—only open questions.”
We believe the most important problems in AI require open collaboration, reproducible research, and tools that belong to everyone. Libertalia exists to advance machine learning through open research.
Our Principles
Open by Default
All research, code, and datasets published under permissive licenses. We believe knowledge should be freely accessible to all.
Reproducible Science
Every paper includes complete training configs and evaluation scripts. Our work can be verified and built upon by anyone.
Community Governed
Research directions decided collectively by active contributors. No single entity controls our agenda.
Quality Over Quantity
We prioritize thorough, impactful research over rushing to publish. Our work should advance the field meaningfully.
Founding Team
Libertalia is led by researchers and engineers with experience across machine learning, security, and startups.
Andre Fu
Co-Founder
Ecliptor (YC W24) Founder, Ex-Twitch MLE, Ex-Microsoft ML, University of Toronto Engineering Science. Published @ CVPR, ICCV.
Egil Karlsen
Co-Founder
VAKTROS Founder/CEO, Former graduate researcher @ University of Toronto, MSc in Computer Science (Cybersecurity & AI/ML), Canadian CTF & Pentesting Champion. Research @ Interac, Okta, eSentire. Ex-CTO of ProSights Labs (YC W24). Published in JNSM and Annals of Telecommunications.
Taha Mohamad
Co-Founder
Chief of Staff @ Deep Tech Fund Advisors, Applied AI Engineer at ProSights (YC'24) & ActivelyAI. Analyst at T-Bird Capital & Republic Capital. CTO at First Generation Investors. Engineer #2 at Compound (acquired). Philosophy & Public Affairs, Claremont McKenna College.
More contributors to be highlighted as we grow.
Current Focus
We're currently working on two major research directions.
Long-Horizon Computer Use
Building and expanding foundational components critical to enabling the advent of long-horizon autonomous computer use tasks.
Aerospace Visual Odometry
Data collection and sensor fusion for visual inertial odometry on dynamical kinetic systems (drones).
Join Us
Whether you want to contribute code, write documentation, review papers, or just learn alongside others—there's a place for you in Libertalia.
Learn How to Contribute