I'm Krisna Wiyana, a SaaS and product designer who spends real time with a problem before I let a single pixel near it. Nine years, forty-plus shipped products, zero shortcuts.
The hard part is sitting with a business long enough to know which screen actually needs to exist, what a user will trust on sight, and where a single decision saves six months of rework. AI can fill a canvas. It can't sit in that discomfort with you.
So I do the unglamorous part first: the questions, the flows, the wrong draftsReal example: for Primal Race I built three full landing-page concepts before picking one. The two we killed weren't bad, they just weren't load-bearing for the actual goal, race-day registrations, not "looks impressive.", before anything gets fast or pretty. What comes out the other side isn't faster. It's correct, and it stays correct after launchAcross 40+ shipped products, the ones that needed a redesign within a year were almost always the ones rushed through discovery, not the ones with weaker visual polish..
Stakeholder calls, support tickets, half-finished specs: I read all of it before I open Figma, because the brief is rarely the real problem.
Wireframes and user flows for the entire journey, so the "exciting" screen isn't built on a structure that collapses three steps later.
Rough, revisable concepts shared early, not a polished reveal, so we find the real disagreements while they're still cheap to fix.
Developer-ready files, states, and edge cases documented, so the slow front-end work doesn't turn into a slow build, too.
The next cards will still be here in eight seconds. Nothing on this page is in a hurry — you don't have to be either.
Drag sideways to browse all five →
Most of these are under NDA, so you get wireframe sketches instead of screenshots. The numbers are real.
A fitness race brand needed a site and design system that matched the intensity of the events themselves, not a generic event template.
Five client accounts, five different brands, each page designed around its own users and goals, not a shared template stretched five ways.
70% of users reorder the same items every 15 days. Designed end-to-end around that habit, cutting purchase steps and lifting GMV.
Dashboards, landing pages, and mobile app designs delivered at a 95% on-time rate through clear prototyping and early alignment.
60% of users wanted integrated payment tracking. Validated the direction, then shipped the MVP in under 3 months.
East Java, Indonesia. Working with teams across Indonesia, the US, and Singapore.
Start a conversation14 drafts, 6 days. Three typography passes, two full rewrites of the hero line, and one intro animation that nearly didn't survive.
1 feature killed. A contact form. It felt faster, but it said less than a plain link.
0 templates. Built with AI, but not generated by it. Every section here went through real reasoning and research before it earned its place — the tool was fast, the thinking wasn't.
2 typefaces. Fraunces for the things that deserve attention. Inter for everything else.
Thanks for taking the long way. — Krisna