About
Understand the system, then improve it.
I like questions that lead to something testable: how a data workflow holds up, what a configuration changes, or which tradeoff makes a system easier to run.
How I work
I start by understanding the input, the intended result, and the constraints in between. Then I build a small, observable path to test the assumptions before treating the first working version as finished.
The same approach carries across my professional data work and my independent local-AI experiments. They are different kinds of work, but both benefit from repeatable tests and notes that make a decision easier to revisit.
What guides the work
- Understand the tradeoff, not just the recommended setting.
- Measure actual behavior before drawing conclusions.
- Document the path so the result can be reproduced later.
- Prefer clear systems that are easier to operate and improve.