At 3 a.m., a smart pet bowl in a customer's kitchen notices it hasn't captured anything in half an hour. The sensor path has quietly wedged — the kind of stall that, on a normal device, would sit there dead until someone noticed the app had
Whenever we build a new website feature, our first instinct is usually to think about a backend. We start planning REST APIs, database tables, authentication, deployment pipelines, and server infrastructure. For many applications, that's exactly the right approach because those components provide the flexibility and scalability needed for
In our previous post we built the first version of an ML bark classifier: 28 features extracted from collar IMU windows, a RandomForest winning a model bakeoff, bark F1 of 0.714 on a held-out test set. That number looked good. It was wrong. The Number That Was Lying to
Search is usually treated as a solved problem. Store the data, index the field you plan to query on, look it up. That holds up perfectly well right until the moment the words someone actually types don't match the words you indexed. The moment exact-match search quietly fails
Every so often, a constraint on a board is so tight it stops being a nuisance and becomes the whole design. This was one of those. We had a carrier board that needed to control the power state of several system-on-modules — up to six of them from one carrier — and
The first minute a customer spends with a connected device decides how they feel about it forever. Unbox a smart pet bowl, open the app, and within about a minute it should just be online — quietly uploading, ready to watch over their pet. But that minute hides a genuine chicken-and-egg
Our IMU-based activity tracker clips to a dog's collar and, in principle, can tell you how far your dog walked today. In practice it needs a teacher. An accelerometer strapped to a moving neck produces a beautiful, noisy waltz of a signal, and turning that waltz into a
Why a part that survives a load test can still fail in the field, and how designing for wear, not just strength, improved the long-term reliability of Everbowl. One of the first questions a mechanical engineer asks about a part is "will it break?" FEA, hand calculations, and
When most people think about the engineering behind a pet wearable, they imagine motion sensors, batteries, AI models, or health analytics. Those are the visible innovations. What they rarely think about is the invisible layer that determines whether any of those innovations reach the user at all: Bluetooth connectivity. A
Before a page shows up in search results, a search engine has to know it exists. That's indexing, and it normally happens through crawling: a bot follows links from pages it already knows about, reads your sitemap, and revisits your site on some schedule it decides for itself