Yongwoo Lee

I believe the process of turning ambiguity into clarity creates value in itself. The judgment that lives only in experts' heads, the decision context buried in an organization's day-to-day work, the conventions everyone knows but no one has written down — without structure, organizations repeat the same mistakes, and AI only mimics the surface. What I want is to design the relationships and judgment factors of the real world — the ones that don't show up in data — into a form AI can actually reason with.

What draws me in is breaking through bottlenecks and making what's tacit precise. Ontology building is exactly that work — turning chaos no one has organized into structure. And it's hard. As Palantir puts it, "pain is our moat." Extracting and structuring an organization's tacit knowledge is a domain no one can easily replicate, and that's why it creates the most value. Easy work gets replaced fast; hard work becomes a moat.

That's why I work on ontology and agent harness design. Ontology captures the judgment and relationships that live only in people's heads — the context that doesn't show up in data — into a form AI can reason with. The harness is the framework that lets agents act on that context reliably. Together, they're the foundation for AI to think like an organization, not just mimic its surface.