Our technology
Edge AI
Models that run on the device itself: cameras, robots, vehicles and industrial PCs, working in real time without an internet connection.
We quantise and optimise models so they fit the compute already in the machine, the vehicle or the camera. Decisions are made on the spot in real time, and the system keeps working when the network is down.
Because data is processed where it is created, images, audio and measurements never leave your site. Each release is regressed on real devices before rollout.
Research on arXiv
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
- Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference
- Deep Compression: Pruning, Trained Quantization and Huffman Coding
Used in
See it on your own data
We start with an MVP on your screens, devices or tickets. Request an MVP and bring a real use case.
Request MVP