What guides us
AI should become more capable without making human intent less visible.
We are building the infrastructure that keeps expert judgment inspectable, executable, and connected to the systems AI engineers improve.
Bring human intent into AI behavior
We are building the infrastructure that lets domain experts and AI engineers shape reliable systems together.
Our researchTreat correctness as an engineering system
Coverage, judgment, evaluation and improvement belong in one connected development workflow.
Our platformBuilt across San Francisco and Tokyo
A multidisciplinary team working at the intersection of AI engineering, domain expertise and human alignment.
Meet the team