TinyML / embedded inference

The practice of running machine-learning models on microcontrollers and other resource-constrained embedded substrates rather than on servers or cloud GPUs.

grounded in: Latest AI/tech trends theme 'On-device / embedded ML: Momentum is toward running models locally and in unexpected substrates rather than always calling the cloud.'

Connected concepts

Small language model, CPU inference serving, On-device speech recognition, Model offloading