AI
The last post ended on an open problem: we could not make phone-photo galleries work well enough to matter. Adding more enrollment images made things worse, not better. This post is about what we changed and what finally moved the numbers. The setup Before getting into results, it helps to
Hardware
Most of the interesting hardware work at Hoomanely isn't greenfield. The connected feeding bowl already had a working carrier board: a compute module doing the thinking, a precision load-cell front end doing the weighing, connectivity, power, the usual. The board did its job. So when two new product
Most SEO conversations start and end with content - better keywords, longer articles, more backlinks. That's a real part of the picture, but it quietly assumes something worth actually verifying: that search engines, and increasingly AI answer engines, can find your pages, read them correctly, and trust what
Software
One message stopped testing cold: "my bowl is not getting connected to wifi." We have an FAQ that answers exactly this. The bot escalated anyway. No answer sent, just a ticket sitting in Slack for someone to pick up six hours later. Nothing was broken, technically. Retrieval ran,
AI
In our previous post we hit around 92% identification accuracy on detected frames. That result came with a catch we had quietly set aside: the gallery was built from bowl camera images, evaluated against bowl camera images. What happens when that changes? The Problem With the Current Setup A quick
When I first started working with Grafana, I treated it like a collection of graphs. I could see request counts, latency, error rates, and dozens of other metrics, but I didn't really understand what they were telling me. Every panel looked useful on its own, yet I found
Hardware
Filter placement, return-path proximity, connector-side filtering, trace length control, and shield-boundary discipline EMI problems are often discovered late. The schematic looks correct. The filter components are selected from the datasheet recommendations. The simulation shows attenuation. The PCB passes functional testing. Then the product enters certification testing, and suddenly a few
AI
Our smart bowl fires its camera whenever anything comes within 40 centimetres — a dog's nose, a human's ankle, a broom. That is by design: in the data-collection phase, every visitor is a labeled training sample. But the product question is sharper than the trigger: the bowl
Product
An engineering investigation into how sensor geometry, optics, and software together determine the true field of view of an integrated camera system. Introduction When you build a vision system out of off-the-shelf parts, it's tempting to assume that if the sensor and lens are mechanically compatible, they'
Technology, Insights and Ideas.
Firmware
Replacing the operating system on a computer usually means having your hands on it: a cable, a USB stick, a recovery button, someone in the room. Now imagine that computer is inside a pet's feeding bowl, in a stranger's kitchen, behind their home router, three time
AI
The last post ended on an open problem: we could not make phone-photo galleries work well enough to matter. Adding more enrollment images made things worse, not better. This post is about what we changed and what finally moved the numbers. The setup Before getting into results, it helps to
Hardware
Most of the interesting hardware work at Hoomanely isn't greenfield. The connected feeding bowl already had a working carrier board: a compute module doing the thinking, a precision load-cell front end doing the weighing, connectivity, power, the usual. The board did its job. So when two new product
Most SEO conversations start and end with content - better keywords, longer articles, more backlinks. That's a real part of the picture, but it quietly assumes something worth actually verifying: that search engines, and increasingly AI answer engines, can find your pages, read them correctly, and trust what
Software
One message stopped testing cold: "my bowl is not getting connected to wifi." We have an FAQ that answers exactly this. The bot escalated anyway. No answer sent, just a ticket sitting in Slack for someone to pick up six hours later. Nothing was broken, technically. Retrieval ran,
AI
In our previous post we hit around 92% identification accuracy on detected frames. That result came with a catch we had quietly set aside: the gallery was built from bowl camera images, evaluated against bowl camera images. What happens when that changes? The Problem With the Current Setup A quick
When I first started working with Grafana, I treated it like a collection of graphs. I could see request counts, latency, error rates, and dozens of other metrics, but I didn't really understand what they were telling me. Every panel looked useful on its own, yet I found
Hardware
Filter placement, return-path proximity, connector-side filtering, trace length control, and shield-boundary discipline EMI problems are often discovered late. The schematic looks correct. The filter components are selected from the datasheet recommendations. The simulation shows attenuation. The PCB passes functional testing. Then the product enters certification testing, and suddenly a few
AI
Our smart bowl fires its camera whenever anything comes within 40 centimetres — a dog's nose, a human's ankle, a broom. That is by design: in the data-collection phase, every visitor is a labeled training sample. But the product question is sharper than the trigger: the bowl
Product
An engineering investigation into how sensor geometry, optics, and software together determine the true field of view of an integrated camera system. Introduction When you build a vision system out of off-the-shelf parts, it's tempting to assume that if the sensor and lens are mechanically compatible, they'