Writing about the things I build, the places I go, and whatever else is on my mind.
Building CineLens, a conversational movie discovery agent that swaps between Claude, Gemini, and local models with one environment variable — and what breaks when the cheaper model makes mistakes.
A dirt fire road under transmission towers at Rancho San Antonio Preserve became the thing that made Sunnyvale feel like home after moving from Texas.
A story about a father who let his daughter enjoy warm pakodas before telling her she'd failed an exam, and what it reveals about love, marks, and withholding.
Every claim the series made, in one place, stated plainly and traced back to the experiment that supports it.
The exact commands, in order, including every failure and fix, for deploying a FastAPI + React app to Cloud Run.
What happens to old models, how labs actually build the next generation, and what's really changing under the hood when they do.
The gap between an application that runs and one that ships, written from an actual deployment rather than a plan for one.
Using small models as evaluators for other small models, and finding out how much you can and can't trust the judge.
A controlled comparison of a base model and its instruction-tuned sibling, testing whether assistant behavior is installed or emergent.
How text becomes tokens, how probability distributions become text, and what happens when you push both to failure on purpose.
Why I built an LLM playground instead of training a toy model, and what the architecture teaches before any experiment runs.