Notes, explanations, and perspectives from learning by doing. Writing forces me to slow down, test what I think I understand, and make the lessons useful beyond the experiment itself.
Retrieval-augmented generation makes much more sense once you stop treating it as a diagram and actually build the pipeline. This article breaks down the moving parts, the assumptions, and the questions I encountered along the way.
What carries forward from enterprise data and analytics, what needs a different mental model, and why AI changes more than just the technology layer.
The projects, mistakes, experiments, and lessons behind documenting a technical transition while it is still happening.
Future writing will follow the actual work: retrieval quality, evaluation, agents, MCP, local models, automation, and whatever breaks next.
These aren’t fixed categories. They reflect the areas where projects, experiments, and prior experience keep intersecting.
The goal isn’t to publish constantly. It’s to document the ideas that become clearer through building, testing, questioning, and writing.