An automated investment product. Users hand real money to a system that acts on their behalf, so every screen has to earn their trust.
The only frontend engineer, working with the founders. I built the whole web application from the first commit and wrote roughly 80% of the codebase. When the initial designs ran out, I took over the design too, and from then on I was solely responsible for the product’s UX and UI.
Complex concepts, real money at stake, and users who need to feel in control of a system that acts for them.
Struck out: Initial designs from an external designer who wasn’t available to carry them further.
Inserted: I finished and extended his work, then designed every new section myself.
I became solely responsible for UX and UI, and the client came to trust my product decisions.
Struck out: Relaunching a strategy needs two separate wallet confirmations, and users often miss the second.
Inserted: A guided dialog with visible, labelled stages and a live status for each, from creating the vault to the deposit landing.
More users stayed in the app. I found the problem in the product’s user community.
Add tooltips to explain the complex features.
Inserted: I wrote them, and proposed a knowledge base: 14 articles in 5 sections, searchable from a help button on every page, with info links that open the right article in context.
The client accepted it, and I built it.
Portfolio-history charts with tax events marked on them.
Inserted: A short video prototype, before any production code.
Approved. The charts grew into a wider data-visualisation layer.
Struck out: A live, step-by-step progress view for a background process that takes several minutes.
Inserted: Postpone it, and ship the higher-priority features first.
It didn’t fit the architecture of the time and would have cost a lot. The team shipped what mattered first.
Each was handled deliberately, not left to a default. I also proposed and set up error monitoring, with reports routed through the app’s own domain so ad-blockers don’t drop them. Bugs became faster to reproduce and fix.
Shipped to a beta community of real users, whose feedback stayed at minor requests. More users stayed after the guided dialog, the knowledge base shipped, and the charts became part of a wider data-visualisation layer.
“He evaluates the client’s request first. If it doesn’t make sense to him, he doesn’t hesitate to point out that it may not be good for the user experience.”