Product architecture, backend, and system integration
Attendora
Automated attendance system combining Node.js APIs, computer vision, and ESP32CAM hardware into a single practical workflow.
- Node.js
- Python
- OpenCV
- ESP32CAM
Context
Attendora started from a simple tension: attendance tracking is often either manual enough to be annoying or automated enough to become brittle. I wanted a system that treated hardware, computer vision, and web infrastructure as one product surface instead of three separate implementation problems.
Build highlights
- A Node.js API handled the orchestration layer and made the system approachable from a product point of view.
- Python and OpenCV took care of the recognition workflow where accuracy and runtime tradeoffs mattered most.
- ESP32CAM hardware kept the project grounded in a real deployment shape rather than a purely simulated pipeline.
Challenge and tradeoffs
The hardest part was not writing the first version. It was deciding where to simplify. Computer vision projects can become demos very quickly if they are not constrained by reliability, hardware limits, and operator experience.
Result
The result is the kind of project I like most: one that crosses software boundaries and forces the architecture to stay honest.