- 10 days · From kickoff to launch of AI Interview Coach, built on the design system
- +6.1%p · Final offer rate for people who used the coach
- 83K · New sign-ups from season 1 of the Salary Lottery
Products I built with the design system
As Wanted grew from a recruiting platform into a career platform, I took on several new products as a product and platform designer. Alongside each product's business goals, I wanted to help the team's goal of raising design quality and speed, so I used the design system heavily in every product I touched.
The bigger the system, the harder it was to use
A design system compounds: the more it's used well, the more it pays off. But every new component is one more thing to learn. A designer shipping next week doesn't want to study guidelines, and it's often unclear which component fits. Drawing it yourself feels faster.
Forcing people to use the system would have backfired. So I decided to apply it to my own products and show, by example, that a system used well makes you faster. I hoped the engineers I worked with would get comfortable with it along the way, and that other makers would have best practices to copy.
Proof 1: AI Interview Coach, shipped in ten days
After I picked up a prototype another team had casually shared in Slack and gave quick feedback, I joined the Interview Coach project as its designer (one lead, one designer, one front-end engineer). It reads a job posting link and gives likely interview questions and feedback on your answers.
Using the design system for almost 100% of the UI, we launched in ten days. It proved its potential, got more entry points such as job detail pages, and raised users' final offer rate by 6.1 percentage points. AI Interview Coach
Proof 2: the Salary Lottery, with 83K sign-ups

In a task force formed around the marketing team's idea, I designed the product flow and UX on the design system. Two UX ideas drove sign-ups at scale.
- Show winners updating in real time on the first screen.
- Let people enter again when they share with a friend.
We far exceeded the goal of mass acquisition: season 1 alone brought 83,000 sign-ups and 170,000 participants. Season 2 added missions to drive activation after sign-up, collect career information and raise DAU. Article (KR)
Proof 3: Career Map, from career questions to real companies
Career Map predicts your future career path from your job family, role and salary. But after seeing their next career path, many users didn't know what to do next and left. I redesigned the flow so they could compare companies and continue to Wanted's company pages.
- New display logic that defines tier-1 and tier-2 companies and mixes them.
- Comparison views on predicted positions: your current company against the prediction on pay, growth, stability and environment.
After launch, a third of Career Map visitors compared multiple companies, and about 17% went on to the Wanted page of a recommended company.
Proof 4: AI Career Mentor, shipped without a PO
Career Mentor is an AI product that listens to a job seeker's background and concerns and gives advice. Users make better career decisions, and Wanted gets qualitative concerns it can't see in metrics as input for new projects.
- In three weeks without marketing, it collected 872 responses, with AI sorting each by category and core concern.
- Even with preset questions available, 31.5% of people typed their own.
There was no PO on the project, but the time the system saved let us ship on schedule.
Where it landed: examples changed how other teams used the system
Wanted's design platform team measures each team's system usage regularly. Sharing these examples steadily helped narrow the gap in system usage across the company.

