
Build a reliable real-time health device using open hardware, open protocols, and open-weight AI
A working health device on a Raspberry Pi, with open-weight AI at the edge, and the production failures you should design around before they happen.
Megha is an AI engineer building an affordable diabetes and fitness companion on the Raspberry Pi, drawing on the open-source diabetes-tech tradition of Nightscout, OpenAPS and the #WeAreNotWaiting community. Megha also built OCTina, an end-to-end open-source deep-learning pipeline for retinal disease classification.
- How to combine open hardware, open protocols and open-weight models in one real-time device
- Which failures appear in production, and how to design for them from day one





