About

I build systems at the intersection of ecological dynamics, information networks, and economic incentives — what I call symbiotic intelligence: the design of agents and platforms that sense, coordinate, and adapt alongside living environments rather than extracting from them.

Oasis-X started as a robotics and sensor platform for indoor agriculture and climate research. The core insight — that food systems, ecosystems, and economies are all complex adaptive networks — turned out to be much more general. The platform has grown into a broader framework for building intelligence that is genuinely entangled with the physical and social world it operates in.

Research

This blog is where I publish research that sits at the boundary of theoretical and applied machine learning. Current threads:

• Ontological Induction — grounding language generation in categorical proof objects to eliminate hallucination at the structural level.
• Sheaf-Theoretic Reward Spaces — applying sheaf cohomology to RLHF feedback to detect and decompose preference inconsistencies; geodesic policy optimization for hard geometric safety constraints.
• Swarm coordination — multi-agent systems with emergent behaviour, applied to distributed sensing and stigmergic task allocation in physical environments.

Work

At Oasis-X I lead product and research. We are building the infrastructure layer for intelligent adaptive systems in agriculture, climate monitoring, and distributed sensing — with a longer arc toward open, ecological computing.

Contact

Reach me at mike@oasis-x.io or find the source code for most of this research at github.com/MikeHLee.