Studio
Quiet intelligence, built for the real world.
An applied machine learning and IoT studio. We build systems that measure something real, run where the work happens, and keep their claims small enough to verify.
How we work
01
Measure first
If we cannot define the number that has to move, we do not start. A model with no metric is a demo, not a product.
02
Run at the edge
Inference belongs near the sensor. It keeps latency honest, keeps data on site, and keeps the system working when the network does not.
03
Small claims, verified
We would rather state a modest result you can reproduce than a striking one you have to trust.
04
Boring where it counts
Novelty in the model, conservatism in everything holding it up: storage, deploys, migrations, backups.
PeopleMeet the team
What we build on12

