We build infrastructure for persistent understanding of the physical world.
Perception has become cheap. Understanding has not.
That gap is the next foundational layer of computing.
Our work begins in national security, where uncertainty is highest and the cost of being wrong is greatest. It does not end there.
Current AI sees. Current AI reasons.
Future AI must remember.
Memory over snapshots.
Calibration over confidence.
Evidence over fluency.
Humans remain accountable.
We’re building difficult things.
The problem is calibrated inference over sparse, unreliable observation of real places — held across months, and honest about its own gaps.
If you’ve worked on probabilistic inference, spatiotemporal modelling, distributed systems, applied mathematics, machine learning, computer vision, robotics or systems engineering, we’d like to hear from you.
Private briefings by request.
abhishek@deephorizon.io