As a general philosophy, the Edge Impulse platform aims to provide a developer the functionality they need, with sensible defaults, for common edge AI tasks and focuses on extensibility to handle the rest. One such example of this is with anomaly detection.
This is a companion discussion topic for the original entry at https://www.edgeimpulse.com/blog/a-convolutional-autoencoder-for-time-series-data-anomaly-detection