I stopped writing here about four years ago. When I look at the dates on the posts below, it feels like another life. A lot has changed since then, both in my career and in the technical landscape around me. This post is my attempt to close that gap, at least a little.

The leap

Four years ago I did something I would have called impossible not long before. I left Red Hat and joined Redpanda. I did not overthink it. I just did it, and it turned out to be the right call.

Redpanda was a streaming startup. When I joined we were around fifty people. For me it felt like going back to the early days of my career, when I was in a small team and we could build anything we had in mind and put it in production the day after. I had spent so long inside Red Hat’s QE process, climbing that tall staircase of reviews and gates, that I had forgotten things could be much simpler.

A real production system

What got me was working on a real production system that we actually owned. Real customers, real production issues, real consequences. Being on call and knowing the whole cloud could land on you, maybe on an otherwise quiet Saturday morning, was terrifying and refreshing at the same time. And then there was the other side of the same coin: you build something, and a couple of days later real people are using it. That loop, from idea to production to someone depending on it, is addictive.

I am curious by nature, and that curiosity pulled me into nearly every corner of the codebase. I rarely stayed in my lane. I also met some of the best people I have worked with, people I will not forget and shared a lot with.

The serverless team

At some point managing infrastructure stopped being enough. I wanted something harder, so I joined the Serverless team. I underestimated the project, badly, at the start. By the end I was grateful for exactly that, because the problems I ran into there were unlike anything I had faced before.

I learned to read the Kafka protocol on the wire to build features and to chase down customer issues. I learned Go concurrency the hard way, the kind of way that leaves marks. I failed more than once. Each time the team picked me back up, and that is why the Serverless team is one of the best teams I have ever been part of.

I have carried impostor syndrome for most of my career. It has shaped a lot of my decisions, often without good reason, and I know that now. But on the serverless team, for once, the feeling was not lying to me. The work really was that hard, I really did struggle, and I came out the other side anyway.

Why I left

So why leave. The world is changing, and Redpanda is changing with it. At some point I felt I needed a different environment, a clean slate, a new problem to throw myself at. None of this is a complaint about Redpanda. It is more that I do my best work when I am slightly out of my depth and building from close to zero, and I felt the pull to do that again. So I am starting somewhere new. More on that another time.

Why I went quiet

And then there is the writing. Four years of nothing.

Part of it is laziness, and I will own that. But part of it is that the work did not lend itself to writing. What I built could help some customers, but most people would have no reason to care, and there was nothing they could go and try on their own, because the code was not public.

That is mostly an excuse, though. Nothing was really stopping me. In all that time I did not even try to submit a talk to a conference. I know writing, like speaking, is closer to a full time job than a hobby, or at least it used to feel that way. I want to change that. I want to fold writing back into the way I study and think, not as an afterthought but as a required step. After enough time spent learning something, you have to write it down. That is the rule I am setting for myself now.

So consider this the first one after a long silence. It is good to be back.