i take on things i technically shouldn't know how to do yet, learn them fast, and produce results that make people double-check whether i actually understand it or just got lucky. it's usually the former.
this pattern has produced real, working output in: web dev, mobile apps, video/photo/audio editing, music production, 3D modeling, UI/UX, robotics, electronics, PCB design, CAD, hardware modding, soldering, game modding, and enough repair work to count as a skill on its own.
take the task → research fast, no hand-holding → build it → fix it until it's solid → move on
| robotics | drones & hardware | code | design & CAD | iot / automation |
|---|---|---|---|---|
| robosoccer | drone building | comp programming | 3D modeling | arduino stuff |
| robowar | flight comps | algo challenges | CAD & sims | raspberry pi |
| robosumo | PCB design | "logic and reasoning" | technical design | smart systems |
also: debates, GK comps, and various other things i agreed to before thinking it through.
commit counts and green squares aren't a proxy for good work — when a metric becomes the target, it stops measuring anything useful. see Goodhart's Law if that needs backing up.
tl;dr: don't evaluate code by the metadata around it.
|
|
- learned cinematography, audio design, and modding through video game camera mods — credit to Matti Hietanen and Mikhael Sharov for the rabbit hole
- fixed printers repeatedly, unwillingly, successfully
- won robotics competitions with robots designed and assembled the night before
- built drones that flew — competition-grade, not just "didn't crash immediately"
- can hold my own arguing politics or debugging code, same energy either way
