Most of my work lives where a product's data has to stay correct across systems that were never designed to agree.
I work across the whole stack, but the part I care about most is the layer underneath: the API surface, the data model, and the integrations that decide whether a product's information stays right as it moves between systems.
Everything I've built has been live, with real users and real consequences when it broke. That shapes how I work. I don't hand a project off after design, or even after deploy. I stay on it through testing and into production, watching for what the monitoring tells me when something's gone wrong.
AI agents are part of how I build now, and I'm deliberate about it. The workflows I design are structured and reviewable, so the speed never costs me understanding of the architecture or the code. I've also worked with teammates and stakeholders on getting there themselves, mostly by setting expectations and building process around tools they already had.
The front end gets the same care. It's the part of the system people actually touch, so it has to be as solid as everything underneath it.