I’m excited to be joining Eventual.

I became a distributed systems engineer by accident. In 2010 I was working on a startup that suddenly got tens of thousands of users overnight, which was wonderful for our fundraising prospects and terrible for our product’s availability. I had to rapidly learn about horizontal scaling, rate limits, queues, backpressure, and database replicas. It was a crazy time, with a lot of interrupted sleep because RabbitMQ was on fire again, but I was hooked. I realized I found distributed systems problems interesting to work on.

A few years later, I joined Honeycomb. Their customers needed to store and analyze large volumes of observability data, and they had built a special-purpose distributed data store from scratch to handle it. Building your own data store is a bold choice for a startup, but I saw how much flexibility and speed it gave customers - and how much better Honeycomb was than the monitoring products of the day - because they’d built a distributed system that matched their product needs. It’s one of the highlights of my career to have presented that system at Strange Loop.

However, not everybody finds distributed systems as interesting as I do. I talk about other things at parties. Even among software engineers, it’s a specialized topic - a long way removed from the actual problems they solve for customers. Distributed systems is often an uncomfortable reality forced on them by having a lot of data.

Working with large volumes of data can be difficult and counterintuitive. Answering even simple questions - “how many people viewed this product yesterday?” - can be expensive and slow, even if you have a suitable way to query it in place already. Worse, even once you can efficiently answer those questions, apparently similar questions - “how many people in the last minute?”, or “how many people from our loyalty program?” - might be even slower, or need custom data engineering projects. It’s both frustrating, and can lead to bad decisions, when something that should take thirty seconds takes three weeks.

So when Sammy, the CEO of Eventual, told me he knew a lot of companies with large data lakes that they struggled to get value from, I started nodding vigorously. When I learned Eventual had built a distributed query engine - letting product engineers build large-scale data processing pipelines with a few lines of code - I got excited. I saw a rare chance to do interesting distributed systems work while actually building a customer-facing product, with the potential for a really big impact.

So I’m joining the really smart folks at Eventual - to build distributed systems so you don’t have to!