Make the decision visible
Good systems explain their boundaries and leave fewer surprises for the next person.
About
Engineer, builder, and careful optimist.
I like ambitious technology best when its tradeoffs are visible and the people using it remain in control.
I’m KahWee Teng, a staff software engineer focused on frontend systems.
I work where product, architecture, and developer experience meet. I care about making large systems understandable—not only fast.
I write when I have something concrete to add: a migration that took longer than promised, an AI workflow that changed how I review code, or a home project that failed for an unexpectedly useful reason.
The articles explain the decisions. KahWee Labs on GitHub keeps the runnable examples, benchmarks, and deployment companions behind them.
The longer path
From Singapore to San Francisco.
I’m originally from Singapore and now based in San Francisco. That distance is part of how I think: technology is never only the system in front of us. It carries assumptions about place, access, incentives, and the people expected to make it work.
This journal is where I slow those assumptions down. I document the decision, show the evidence I have, and leave enough context for someone else to disagree usefully.
Operating principles
How I work
A good technical decision should remain understandable after the meeting, the migration, and the person who proposed it.
Build the useful layer
The best platform work makes many product decisions easier without hiding their cost.
Write from the work
I prefer tested experience, measured tradeoffs, and examples that survive contact with reality.
Current field notes
What I’m exploring now
Questions that are active in the journal, not claims that have already settled.
AI as an engineering constraint
How model choice, quotas, context, and review practices are becoming ordinary parts of software architecture.
Read the AI practice seriesWhat makes migrations worth it
Where upgrades create durable leverage—and where they simply move complexity into a new shape.
Read the migration seriesInfrastructure at human scale
Self-hosting and home systems as compact laboratories for reliability, privacy, and interoperability.
Read the home systems series