i was in design school when facebook happened. i taught myself to code to build what i could only mock up. now i build the part nobody designs: the infrastructure ai needs to make sense of the mess. and i show leadership teams how to build the same.
Currently Building
shipping
vode(opens in a new tab)
on-device dictation for mac. you speak, it types.
lekto
competence assurance. who can do what, on the record.
ain
workout app. every training plan, structured.
denta
the data layer dentistry never had.
Zzz
context-aware app for baby sleep. AI-first.
Mission
garbage in, garbage out.
most of the data we generate is chaos. recipes without structure, fitness logs in silos, apps that can't share a line of context. and when an ai tries to help? it hallucinates. i build the infrastructure layer, structured apis, unified schemas, data an ai actually understands. i build this myself, every day. that's how i know what runs in a company. and what only runs in a demo.
Research
the rabbit holes
| Title | |
|---|---|
| no api, no intelligence the age of ai is an age of apis. an intelligence is only as smart as the data it can read, and dentistry runs on software from a time that never asked the question. | |
| cognitive retrieval systems exhausted parents don't need hallucinated advice at 3am. a retrieval system that knows when to answer, when to ask, and when to say "i don't know." | |
| athletic data protocol you save workouts from ten different creators. none of them talk to each other. a system that reads structure out of any workout video and makes it comparable and trackable. | |
| verifiable health intelligence personalized health recommendations. your data never leaves your device. not a privacy policy. math. | |
| meal metrics ai grandma's recipes don't come with nutrition labels. an api that structures any recipe and calculates macros. |