Hello World

I’ve been meaning to start a blog for a few years. The usual excuses applied: not enough time, nothing interesting enough to say, someone smarter has already written about this.

What changed is that I kept noticing a gap between the conversations I was having with coworkers and the things I was publishing online. The interesting stuff — the tradeoffs we actually argued about, the things that didn’t work and why, the mental models that took months to build — stayed inside the company. My public presence was just a LinkedIn profile and a GitHub with green squares.

Writing publicly forces a kind of honesty. When you’re explaining something to a coworker, you can wave your hands and say “you know what I mean.” When you’re writing for someone you’ve never met, you have to actually know what you mean. I want that discipline.

Most of what I’ll write about is what I work on: machine learning infrastructure, distributed systems, reliability engineering. Some of it will be broader — career stuff, things I’m reading, ideas I’m working through. I’m not planning to optimize for traffic or algorithmic discovery. I just want a place to think out loud.

If something here is useful to you, that’s great. If you want to push back on something, email me.