@ 외주 턴키AI Career Diagnosis App for Retirees
MVP shipped in 2 months (solo full-stack). Validated: 4.6 on Google Play, 1,000+ downloads.
A contract where what stayed afterward was not a person — it was a system.
── The mission as given ──
“Build the design system and Storybook component library for Samsung Fire's direct service.”
A turnkey schedule, a single full-stack person, a one-off engagement that ended when the calendar ran out. I could have just shipped a handful of components and walked away.
If it ended that way, someone would have to come back the same way every time a new page was added. The real ROI of the contract was a system staying behind, not a person — that recognition was the starting line for the whole job.
Field notes
A one-shot turnkey contract
When the schedule ends, what stays in the client's hands is just the library.
Solo full-stack
The same person runs both component coding and page application.
Built Storybook alongside the code
A way to hand off to the next person through documentation, not raw code.
Shaped common UI — Button, Form, Modal — into variant + props. The next developer can reuse them by changing props alone.
Documented props, variants, and per-state cases as a visual catalog. Handed off to the next person through docs, not raw code.
Did not just build the library — applied it directly onto live service pages. Cases that misaligned during application went back into the library.
tradeoffClean library abstractions vs. the variety of real service requirements

New-page build time
−50%+
Assembled from the library
Storybook catalog
Built
Props and variants documented
Designer ↔ developer
A shared language
Visual sharing through Storybook
After the contract
An asset that stayed
The next page still uses the same library
“An engineer who, even on a one-shot contract, codes so a system remains after the person is gone.”
Reframed the brief from "ship a one-off set of components" into "leave behind a reusable library."
Modular components, a Storybook catalog, and a cycle of applying them straight into the live service.
New-page build time cut by 50%+. The next developer could work off the library alone.
More work