Hands-on experiments · security & knowledge systems

AI Exploration

Explore AI by building with it—and by examining where it fails.

Learning through building

My AI work spans hands-on experiments, local-model infrastructure, assistant and workbench prototypes, and workflow automation. I’m interested in how these tools can make security work and personal productivity more useful, inspectable, and dependable.

Knowledge that keeps its context

I explore retrieval, graph-based memory, and persistent knowledge systems: ways to connect an answer to its sources and preserve useful context across tasks. Data sovereignty and local control are part of the architectural question.

Testing behavior and judgment

My experiments also examine classifier behavior, AI safety and governance, and the gap between a plausible response and reliable work. The questions are practical: what evidence supports an output, how can a person challenge it, and where does the system fail?

Ongoing exploration

This is active development and research across several ideas and tools. I share individual experiments on their observed merits, with scope and limitations kept visible.