Work — YOLANI OLIVAS

Rebuilding a loan marketplace

Shopping for a loan on Credit Karma meant scrolling one long list of hundreds of look-alike offers. The marketplace was missing its revenue goals, and it was so fragile that the team had learned not to touch it.
Redesigned it from a flat list into a grouped feed: the best offers pulled up front, clear sections, and a tighter offer card with fees made visible. Proved the idea on a smaller audience first, then scaled it, and held steady when early results dipped for reasons that turned out not to be the design.
The new structure let smarter offer ranking build on top, driving a multi-million-dollar monthly revenue lift. The first release alone lifted revenue per user by 5%.

Members faced hundreds of near-identical offers, and the filter that was supposed to help was easy to miss, full of options nobody cared about, and gave no feedback that it was working.
Started with research, not screens: five different filtering concepts tested with real loan shoppers. Then rebuilt filtering around how people actually decide, with the filters that matter up front and a live count of offers as you narrow.
The overall numbers looked flat, but the deeper read told the real story: heavy use of the new filters and clear wins for members with lots of offers. The team kept it, and every filter tap now helps rank offers better.

Talking to users kept losing out to deadlines, because setting up a study took days of manual work. So research got skipped, and the team designed on guesswork.
Built an AI-powered system of five reusable workflows that turn a vague idea into a study plan, participant questions, interview scripts, and synthesized insights in minutes. Then ran the rollout like a product launch: training, office hours, and support.
Days of setup became minutes, which flipped the math: checking became easier than guessing. The whole team got faster, not just one project.