About

I'm JP, and I'm building the Zenith Safety Sensor: a ceiling-mounted sensor that notices when someone may need help — a fall-like event, a cry, breaking glass — and sends an alert. The judgment happens on a small AI chip inside the device. By default, no video leaves it. In its strictest mode, there is nothing to watch at all: no live view exists.

It's aimed first at elderly safety — people aging at home, the adult children who worry about them, and eventually the care organizations trying to look after more residents with fewer hands. A child-safety mode on the same hardware comes later.

This is a solo side project, built in the open at whatever pace evenings allow. The log documents the real engineering: benchmarks, false-alarm counts, wrong turns, and the pass/fail criteria I set before running the experiments.

Two honest sentences you should hold me to:

1. The sensor gives awareness sooner — it does not prevent falls, and it is not a medical device.
2. Every privacy claim here is a data-flow fact you can verify, not a security promise. When hardware exists: packet-capture it.

Contact: coming with the email list. For now, the feed is the follow button.