TinyML: Bringing machine learning to the edge
Thoughtworks Technology Podcast
00:00
How to Train a Model and Deploy It on a Device
There's two key personas that sort of I've seen in this area. One where there's like an ML person who's a bit more like me, who comes from the ML side of things and is interested in the embedded. And then the other one's the embedded engineer, who knows how to actually write some bizarre C++ code to get this working because it's complete Wild West. So those embedded engineers who are very specialized in that type of area don't know much about machine learning. It's funny to see a lot of people we work with that are embedded running these ridiculously large models that are like 100 times bigger than they need to be just because it was the simplest thing for
Transcript
Play full episode
