Turning complex problems
into reliable software.

I explore AI agents, developer tools, and reliable software systems—turning ambiguous problems into outcomes that can be verified.

Work samples

Selected projects

Research prototype2026

An experimental platform for verification, recovery, and observability in multi-step AI agents.

In progress2026

A searchable bilingual knowledge base connecting reading notes, code experiments, and long-running themes.

Personal infrastructure2025—2026

A reproducible and reversible Linux workstation and developer-tool configuration.

Field notes

2026.09.08Engineering6 min read

Why exit code zero does not mean the job is done, and how to define meaningful acceptance rules for automation.

2026.08.24Web development5 min read

A language switcher is not enough: routes, content identity, and missing translations all need explicit rules.

2026.08.03Career4 min read

Problems, constraints, tradeoffs, and evidence reveal more than a wall of framework badges.

Have a problem worth solving well?