Working principle
Ship what is real
I would rather show a smaller system that is grounded in code and verification than inflate a roadmap into a fake success story.
About Ethan
I like building where messy workflows, technical curiosity, and practical outcomes all overlap.
I like building software in places where the workflow is messy and the payoff from clarity is high. That includes internal operations tools, AI-assisted reporting, desktop automation, and systems-level customization.
A lot of my work starts from practical friction: duplicated spreadsheets, scattered admin access, slow manual handoffs, or a desktop setup that feels powerful but fragile. I enjoy turning those into something centralized, documented, and easier to evolve.
AI fits into that process in two ways for me. Sometimes it is part of the feature itself, and sometimes it is the multiplier that helps me understand, scope, and ship faster. I care a lot about keeping that distinction honest.
I have also been building a portable bootstrapper for my own AI workflow. It keeps plans, implementation history, and shared agent rules close to the repo so I can manage ideas better and give future AI sessions a concrete record of what has already been tried, changed, and verified.
Working principle
Trace before changing
I like understanding how the whole path connects before I touch shared behavior. That habit shows up in the governed workflow I use across agents and in the way I document implementation decisions.
Working principle
Re-scope based on evidence
If a local-LLM architecture is too slow or a project balloons in scope, I would rather pivot with intent than force the original plan past the point where it still makes sense.
What I want this site to show
AI matters to me most when it increases capability without lowering standards.
Sometimes that means using AI directly in the product. Sometimes it means using it to manage architecture, explore options, or compress the learning curve on a harder system. In either case, I still care about edge cases, truthful labeling, and whether the result genuinely works.
Hiring-manager takeaway
- I can ship product-facing AI features without hiding the fallback path.
- I can use AI to keep complex internal tools organized and documented.
- I can go deep into systems work when the project demands Linux, desktop, or runtime-level understanding.
- I am comfortable changing direction when the technical evidence says the original plan is no longer the right one.