AI Science
Edge AI for instruments that have to be right
DCI is Distributed Compound Intelligence: a small-system AI architecture built for constrained hardware, where a single confident answer is not enough. BLEMO is the companion runtime layer that carries that stack onto the instruments we actually ship to the bench and the barn.
Why it exists
Most edge models return one label and move on. Farm and veterinary instruments cannot afford silent confidence. DCI is designed so compound pathways can challenge each other before a result is treated as settled. BLEMO is how that architecture rides on real boards next to real optics and microfluidics.
Verification minded inference for niche scientific domains. Built for edge devices, not for a distant GPU farm that the barn cannot reach.
The layer that seats DCI on the hardware we control: cameras, flow cells, local storage, and the operator interface that has to survive a wet lab.
First instrument
FloatScan is the first public instrument built around this stack: veterinary microscopy for fecal flotation and related sample work, with AI assist on device. The science of the sample and the science of the model have to travel together.
Development status
Register last reviewed .
Get updates
Leave an email if you want notice when DCI / BLEMO documentation and FloatScan runs open. Nothing is for sale on this page.