Tags

Working with language models in practice: extracting structured data from ugly documents, building agents, AI-assisted development, and explaining how the models work.

Product decisions from things I ship myself: what to build, what to leave out, how to say what a tool does, and what the numbers say afterwards.

The products I build and run on my own: what each one is, the engineering decisions behind it, and what happened after launch. Every post here links to the live product.

Longer takes on how we work: burnout, AI tooling, expectations, and what I think we should do about them.

TypeScript across the whole stack: React frontends, tRPC and Drizzle backends, pure-function game engines, and the patterns that make solo projects maintainable.

On burnout in software teams: what causes it, what it did to me, and what helped.

Notes from my own career as a software engineer: how I look for work, what interviews are like now, and how the job market has changed.

Building browser games: deterministic simulations, solvers that prove a puzzle is fair, rollback netcode, and the product side of a daily game.

Technical interviews from the candidate side: what they ask, how I prepare, and what actually made a difference.

How I search for a developer job: the process, the numbers, the interviews, and what I would do differently.

How large language models work and how to use them in real systems: next-word prediction, attention, RAG, agents, and the trade-offs of putting them in production.

Collecting data from official sources that were never meant to be read by a program: HTML, PDFs, gazettes, and the pipelines that turn them into something you can query.

How I actually work day to day: tools, AI-assisted development, and the habits that survived contact with real projects.