PortfolioYikkan
Yikkan — Attendance with facial recognition
Production attendance: facial capture, REST API, and PostgreSQL, with containerized infra ready for on-site or cloud.
The project
Yikkan is an attendance control system with facial recognition: the team clocks in/out without card friction, backed by a first-party API and database.
The problem
Reliable attendance in the field or office often depends on rigid hardware or spreadsheets. The need was a operable web product with practical biometric identity and reproducible deploy (dev and prod).
Architecture (high level)
- Backend Django 5 + Django REST Framework.
- PostgreSQL as source of truth for users, records, and media.
- Facial recognition via
face-recognition/ dlib on captured images. - Docker Compose (dev/prod) + Fly.io deploy; Nginx where needed.
- Quality: Playwright for critical UI flows.
My contribution
- Product design and Django/DRF + PostgreSQL stack.
- Facial recognition pipeline and media storage.
- Containerization, environments, and cloud deploy.
- Automated smoke/UI tests.
Visible outcomes
- Browser-usable product with biometric clock-in.
- Infra ready to reproduce on Docker or Fly.
- Foundation to grow HR modules without changing platforms.
Stack (examples, not a ceiling)
Django 5 · DRF · PostgreSQL · face-recognition · dlib · Docker · Fly.io · Nginx · Playwright.