Case study 5
The Recruiter App
The same matching, turned around: many candidates, one job, fair scores.
- Product
- Architect
- Build

Project record
- Role
- Product · Architect · Build
- When
- Sept 2026 to today
- Status
- Built; end-to-end testing in progress
- Built with
- The Job Bank database, rules-based scoring, the AI Airlock, PocketBase
- Open it
The problem
Recruiters screen fast and inconsistently, and good candidates get missed.
What I did
I reversed the job-seeker engine. Candidates are compared against one job with the same keyword scoring, sorted into Best Match, Worth a Look and Weak Fit, and each strong candidate gets a deeper, person-reviewed look.
What shipped
A recruiter portal with candidate profiles, panel interview guides and scorecards, sharing one database with the job-seeker side.
What I learned
Building both sides on one engine keeps the logic honest: the same rules judge the job and the candidate.
Open it live
See it running, or read the plan I worked from.