AI
When you're building a thermal-detection system to watch over a dog, the sensor you pick is only half the story. The lens sitting in front of it — and specifically how wide it sees — quietly decides how far away you can read your pup and how forgiving the aim
Hardware
Crystal tolerance budgeting, PLL lock monitoring, fallback oscillator paths, and clock-valid gating Clock failures rarely look like clock failures at first. A sensor stops responding after a cold boot. A camera interface works on one unit and fails on another. A communication link becomes unstable only after the product warms
Firmware
Two sensors share one I²C bus on our hub. One is an MLX90640 thermal imager sampling an 8×8 temperature grid twice a second. The other is a combination proximity-and-barometer that checks the room distance and ambient pressure at roughly the same cadence. Each sensor looks harmless in isolation.
Every software product begins with a simple vision. The first few features are straightforward, the codebase is easy to understand, and engineering decisions are made with confidence because the system is still small. As products mature, however, the landscape changes. New capabilities are added, teams expand, delivery timelines become tighter,
Firmware
A pet tracker has to be honest about two very different things at once. Step counts come from a continuous, high-frequency motion stream that never stops, while audio arrives in big, bursty clips only when something happens. Both are health data, both must reach the cloud intact — and they have
Product
How Real-World Use Drove Mechanical Design Decisions No matter how carefully a product is engineered, its first encounter with real users almost always reveals something the design team did not anticipate. Engineers interact with products differently than users. During development, devices are handled deliberately, assembled with care, and operated according
AI
In our first post on the bark detection pipeline, we showed how a single threshold on jerk_3d_max gives a useful signal for detecting barks from a dog collar IMU. For any capture window where jerk_3d_max crosses 6.5 g/s, the rule calls it a bark.
Hardware
Why pin assignment is not just routing convenience Dense connectors look harmless on a schematic. A row of pins. A few high-speed lanes. Some power. Some ground. A few control signals. Maybe an I2C bus, reset, interrupt, enable lines, and one or two spare pins. Everything fits. Then the board
Firmware
A pet's day has a soundtrack, and most of it is health data. The rhythmic crunch of kibble, a gulp of water, a bark at the doorbell, the wet cough that wasn't there last week — each is a signal a vet would want to know about.
Technology, Insights and Ideas.
AI
In our previous post we described the multi-stage vision pipeline we built to identify dogs by their facial landmarks - covering preprocessing, embedding model training, gallery building, and real-time inference. This post covers what happened when we ran it on real camera footage and reviewed every result manually. Setup A
AI
When you're building a thermal-detection system to watch over a dog, the sensor you pick is only half the story. The lens sitting in front of it — and specifically how wide it sees — quietly decides how far away you can read your pup and how forgiving the aim
Hardware
Crystal tolerance budgeting, PLL lock monitoring, fallback oscillator paths, and clock-valid gating Clock failures rarely look like clock failures at first. A sensor stops responding after a cold boot. A camera interface works on one unit and fails on another. A communication link becomes unstable only after the product warms
Firmware
Two sensors share one I²C bus on our hub. One is an MLX90640 thermal imager sampling an 8×8 temperature grid twice a second. The other is a combination proximity-and-barometer that checks the room distance and ambient pressure at roughly the same cadence. Each sensor looks harmless in isolation.
Every software product begins with a simple vision. The first few features are straightforward, the codebase is easy to understand, and engineering decisions are made with confidence because the system is still small. As products mature, however, the landscape changes. New capabilities are added, teams expand, delivery timelines become tighter,
Firmware
A pet tracker has to be honest about two very different things at once. Step counts come from a continuous, high-frequency motion stream that never stops, while audio arrives in big, bursty clips only when something happens. Both are health data, both must reach the cloud intact — and they have
Product
How Real-World Use Drove Mechanical Design Decisions No matter how carefully a product is engineered, its first encounter with real users almost always reveals something the design team did not anticipate. Engineers interact with products differently than users. During development, devices are handled deliberately, assembled with care, and operated according
AI
In our first post on the bark detection pipeline, we showed how a single threshold on jerk_3d_max gives a useful signal for detecting barks from a dog collar IMU. For any capture window where jerk_3d_max crosses 6.5 g/s, the rule calls it a bark.
Hardware
Why pin assignment is not just routing convenience Dense connectors look harmless on a schematic. A row of pins. A few high-speed lanes. Some power. Some ground. A few control signals. Maybe an I2C bus, reset, interrupt, enable lines, and one or two spare pins. Everything fits. Then the board
Firmware
A pet's day has a soundtrack, and most of it is health data. The rhythmic crunch of kibble, a gulp of water, a bark at the doorbell, the wet cough that wasn't there last week — each is a signal a vet would want to know about.