Kit intro ML edge BCUB3

An introduction kit to embedded machine learning (edge ML): a compute board (ESP32-S3, Raspberry Pi or Jetson depending on the level) and a path to go from sensor data collection to a small model running locally, without cloud. It makes tangible the central edge trade-off: how much compute power for which model, at what latency and what consumption.

It is a pedagogical on-ramp, not a production platform. It teaches how to choose the right hardware tier — microcontroller for TinyML, Linux SBC for a medium model, Jetson for deep vision — a trade-off developed in the edge guide and on the ESP32-S3 / RPi 5 / Jetson entries.