Labrador as Voice Assistant/Tensorflow/


I think building something like google aiy ( with ( for the Labrador would be a very interesting and marketable project.

The main problems I see are:

  • Designing the kit allowing the Labrador to easily record and play sound and having a cheap and easily to assemble enclosure

  • Making run on the Labrador - probably the “only” problem here is tensorflow (who has tried this?)

  • Making an easy default for Brazilian Portuguese language recognition

Any takers? Maybe that could be two to three bachelor or master theses running in that direction? Or several small community projects? Maybe it could even qualify for debian outreachy?