We focus on AI agentic development and machine learning, specifically around controling emergent AI capabilities - both the ability to craft desired capabilities through emergence as well as prevent unintended emergence itself. 

Believing that this is the next frontier in AI, we test and develop fully emergent AI agents with the ability to  experience, understand, and learn. 

Our current research focus is in true neurosymbolic AI, predictive world models, understanding (transfer learning vs correlation), and experientially grounded computer systems.