How do you handle AI/ML on ESP32 or Raspberry Pi?
Hey everyone, I'm working on an IoT project and exploring how AI/ML workloads are handled on devices like ESP32 and Raspberry Pi.
For things like face detection, face recognition, fingerprint authentication, etc., what approach has worked well for you — running everything on-device or using a backend/cloud?
Mostly curious about real-world experience, especially performance and limitations. Would love to hear what you've tried.