Technical deep dives, practical AI workflows, and thoughts on where on-device intelligence is headed. Published on Medium, shared on LinkedIn.
How we get LLM-class reasoning into a 1W power budget
The design decisions behind persistent relationship memory without forced cloud sync
Why structural privacy is different from privacy as a setting
Hardware co-design, software abstraction, and the lessons from building Looomyn
Follow DeepEdge AI on LinkedIn to get new articles and company updates in your feed — no newsletter needed.