Behind BitPanic

I'm a technical engineer who enjoys solving practical problems. Over the past decade I've worked across machine vision and automated optical inspection (AOI), industrial automation, optics, software integration, technical support, and customer evaluations - from designing inspection strategies that find micron-scale defects to recovering failed production systems and helping people understand why a problem exists in the first place.

Most of these projects come from real problems in machine vision and metrology: getting a camera into perfect focus with as few frames as possible, matching colour across many cameras, stacking and stitching microscope images, or designing calibration targets down to the micrometre. Others are small embedded instruments that read a sensor and put a clear number on a screen.

The common thread is turning a fiddly, manual task into something fast, repeatable and pleasant to use - usually with a responsive GUI in front and careful image or signal processing underneath. What motivates me is not technology itself; it's using technology to solve real-world problems.

It is not just this website. All of the software on BitPanic - and the projects on my GitHub - was created with AI. The in-browser demos, the tools, the image-processing experiments, and this site itself were all written in collaboration with an AI coding agent. I describe what I want in plain language, the agent writes, tests and deploys the code, and we iterate together until it is right. The illustrations are AI-generated, the explanatory diagrams are drawn in code by the agent, and the writing was shaped from my own notes and conversations.

I am genuinely excited about how powerful this has become. It is not about replacing engineering judgement - it is about removing friction. Ideas that used to sit in a notebook for months now become working software in an afternoon. I treat the AI as a tireless pair-programmer and a research partner: it handles the boilerplate, surfaces options I had not considered, and lets me spend my attention on the parts that actually need an engineer.

For me this is also a way to become a better engineer. Explaining a problem clearly enough for an agent to solve forces the same discipline as good engineering: understand the requirement, name the constraints, and test the result. The faster feedback loop means I try more approaches, see more failure modes, and learn quicker. That is exactly the philosophy this notebook is about.