18 / How I Maintain Open Source
I wrote a maintainer's guide for projects that didn't need one yet. Then I opened GitHub and found two pull requests waiting.
Thoughts on software systems, architectural decisions, experiments, cross-project comparisons, and reflections on the engineering process.
I wrote a maintainer's guide for projects that didn't need one yet. Then I opened GitHub and found two pull requests waiting.
A routine cleanup step, unsupervised filesystem access, and the night an AI agent sweep almost wiped my active work.
AI didn't change how I work. It removed the only reason I'd ever need to work differently.
On long-term ambition, foundational software, and accepting the uncertainty of the path ahead.
The workflow, discipline, and practical engineering habits behind how I approach building software.
The path from learning basic programming logic to designing distributed services and containerized evaluation harnesses.
The disciplines, visual craft, physical habits, and philosophy that shape how I approach building software.
A reflection on curiosity, problem-solving, building things from first principles, and why turning abstract ideas into working software felt like the right path.
What I Learned From Building the Same Distributed Cache in Java and Python
From Context Unification to Agent Evaluation
Two Ways of Making AI Systems Less Black-Box
Why I Kept the Intelligence Close to the Application
What Changes When the AI Stops Being a Single API Call
When “The Model Answered” Isn’t Enough: Designing for Failure
The Evolution of My Real-Time Architecture
How My Projects Became More Serious
Why Some of My Projects Need More Than One Backend
What Building for Production Changed About How I Build