⚠️ Screen too narrow

best viewed in landscape or larger window

Embedded AI - Reimagined

Needed a way to bring local AI model to a small project.
Reimagined project for embedded AI to run local AI/ML model w/ pipeline to cloud based servers for training.

Summary of Video:

Reimagine-embedded ai
Reimagine Embedded AI
Reimagine-embedded ai
Problem Statement
Reimagine-embedded ai
Needed AI model to run on an embedded board
Model was way too big to train on the embedded board
Did not want to add RAM and CPU power to battery powered device
Reimagine-embedded ai
Reimagine
Reimagine-embedded ai
Reimagine - Design
Embedded board links to backend platform
Platform does the training
Embedded board runs the AI/ML model
Reimagine-embedded ai
Realize model training can happen on a larger CPU
_ cloud or connected laptop
Created pipeline to cloud
Training takes place on a cloud based server
Model is then sent to an embedded board
_ Tensor Flow lite and TinyML
Embedded board can run the model
Reimagine-embedded ai
Outcome
Reimagine-embedded ai
Project works great and can be battery powered after model is downloaded
Reimagine-embedded ai
Lessons Learned
Reimagine-embedded ai
Training model on a cloud server is the way to go
Creating the infrastructure to this creates superior product
Reimagine-embedded ai
Thank-You