I like problems that sit in the awkward space between disciplines β where a clean software answer isn't enough and you have to think about hardware, biology, or economics too. Most of what I build starts with the same question: what if this thing could just take care of itself?
Most days I'm reading papers, breaking soldering joints, or arguing with a model that refuses to converge.
01. ship the ugly version first
02. hardware will betray you β plan for it
03. the boring solution usually wins
04. if you can't explain it on a napkin, rebuild it



