Hey DEV community 👋 I'm Todd — an AI engineer based in the UK. I spend most of my time building systems that actually use AI rather than talking about using AI. What I work on Claude/LLM integrations — wiring Claude into real engineering workflows. Co-authoring code, automated pipelines, using it where it genuinely adds signal rather than noise. On-device AI — computer vision models that run on iOS with no internet connection. Edge inference, model size tradeoffs, the gap between lab accuracy and field accuracy. Developer tooling — zero-config test runners, CI pipelines with AI in the loop, the kind of boring-but-important stuff that makes teams faster. Personal AI infrastructure — building a persistent AI assistant that runs 24/7, has real memory, and can actually take actions. Not a chatbot demo. Why I'm here I write about the engineering decisions, the tradeoffs, and the things that didn't work as well as the things that did. Practical stuff, real code, honest takes.…