Not every application can tolerate the latency, bandwidth cost, or privacy exposure of cloud round-trips for every inference. Edge AI puts intelligence directly on the microcontroller.

An end-to-end, production-ready TinyML pipeline to simultaneously estimate battery State of Charge (SoC) and State of Health (SoH) on low-power microcontrollers. Features Quantization-Aware Training (QAT) to compress a deep regression model down to 6.2 KB for bare-metal C++ BMS/EV edge hardware deployment.

Complete flow for keyword spotting on microcontrollers. From data collection to data preparation to training and deployment.

Implementation of Hasaki (Neural network CLI tool in C++ with configurable layers and activation functions)

Inference at the Edge. Train on desktop, deploy as pure C.