Smart Attendance System.
A full-stack, hardware-ready school attendance system combining RFID-based identity, 1:1 face verification, configurable attendance sessions, role-based staff workflows, device integration, and derived reporting through one centralized attendance model.
- Atomic and idempotent attendance persistence
- Authenticated Device API designed for real RFID and camera hardware
- Server-side RBAC for system admin, homeroom teacher, and operator roles
- RFID identity with 1:1 face verification rather than broad face search
- Private FastAPI biometric service using YuNet and SFace
- Automated web and Python quality gates with Vitest and pytest
Building attendance around verified identity.
I developed Smart Attendance System as a full-stack school attendance platform that combines RFID-based identity, 1:1 face verification, configurable attendance sessions, staff workflows, device integration, and derived reporting. My work covered the product logic, system architecture, web application, Device API, database and authorization model, biometric verification service, administrative tools, homeroom workflows, reporting, and automated testing.
Making attendance trustworthy, not just recordable.
Scanning an RFID card can identify who is expected to attend, but it cannot prove who is actually holding the card. The system also needed to evaluate face verification, active attendance sessions, participant eligibility, arrival timing, duplicate scans, and staff permissions before an attempt could become a valid attendance record. I treated these as system-level rules rather than interface-specific logic, so the same attendance decisions remain consistent across the public simulator, staff workflows, and hardware-facing device integration.
Identify the card. Verify the person. Trust the record.
Keeping attendance truth on the server.
I separated device input, identity resolution, face verification, attendance rules, persistence, and reporting so no browser screen or hardware client could decide attendance on its own. RFID resolves the expected student, face verification confirms that identity when required, and the attendance engine evaluates the active session, participant eligibility, timing, duplicate scans, and verification result before creating a canonical attendance record. Raw device events, verification attempts, accepted attendance, teacher confirmations, and reporting data remain separate so the system stays auditable and easier to maintain.
One attendance engine behind every entry point.
The public simulator and hardware-facing Device API enter the system through different paths, but both depend on the same attendance rules and data boundaries. This keeps device integration separate from the logic that decides whether an attendance attempt is valid, while allowing verification, persistence, staff workflows, and reporting to stay consistent across the product.
Designed across every part of the attendance workflow.
I designed separate interfaces for verification, administration, student identity management, schedule configuration, homeroom reconciliation, and reporting while keeping the underlying attendance decisions and authorization rules centralized.
Built beyond the attendance screen.
The system connects identity, biometrics, scheduling, authorization, device communication, attendance rules, and reporting as one operational product.
From verification to a trusted attendance record.
Smart Attendance System brings RFID identity, 1:1 face verification, configurable attendance sessions, staff workflows, hardware-ready device communication, and derived reporting into one consistent attendance model.
Try the live demo↗