Employers hire on fit. Candidates get found.
A candidate-led recruitment marketplace. People build one rich profile, employers hire on what they can see of the person, and no recruiter sits in the middle charging a margin. I redesigned the existing product across both portals and designed the capabilities built on top of it.
Recruitment runs backwards, and everyone pays for it.
Candidates apply into a void. Employers judge look-alike résumés and pay a recruiter margin for the privilege of guessing. The poor-fit hires that follow burn out, and the cycle restarts.
And the product itself was working against them.
A platform already existed. The problem was that finishing anything on it took too long, so most people never did: fewer than half of the users who started signing up, building a profile or posting a job reached the end. Three screens accounted for most of that.



Two portals, one direction of travel.
Candidates build one profile that does the chasing. Employers watch a recorded interview and hire from a free applicant tracker. Both portals run on the same component library, so a change lands on both sides at once.
designed
Four surfaces that make the marketplace work.
The redesign fixed what already existed. These four are the capabilities built on top of it, and the reason a candidate now gets found.
A live two-sided product on one system.
Both portals ship from one shared design system: 20,000 candidates, 400 employers and five enterprise clients on it. Completion across the core funnel roughly doubled and the time to get through it fell by about two thirds.
Supplied by the founder. The redesign shipped alongside significant new capability while the platform was growing, so the improvement reflects both the design work and the expanded product rather than the redesign alone.