
Scaling engine for PostgreSQL
Hey! PgDog is a tiny database startup in SF. We’re scaling the world’s most popular database, PostgreSQL. PgDog is in production, used by dozens of companies, large and small. The problem is big, but so is the opportunity, and we’re hiring!
The jack of all trades. Rust, Postgres, infrastructure (you don’t mind Kubernetes too much), networking, distributed systems, and so much more. You’ll build features that will go directly into our customers hands, the day you ship them.
With great power comes great responsibility, of course. You’re meticulous, detail-oriented and are not afraid to take calculated risks. You write tests because you want to make sure the thing works, not to check a box. You’re not afraid to run code in production, on real data; after all, that’s the only way to know something works.
You’ll fix a bug or ship a cool feature. We’ll give it to one of our users or customers to try right away.
You’ll get acquainted with pretty much all of our code. You’ll start coming up with your own bug fixes and features.
You’re designing our roadmap. You’re contributing to everything we’re doing: coding, talking with customers, and even finding us more! Things we badly want to build:
If any of this sounds super hard…well, that’s because it is, on the surface. We’re splitting this up into small, achievable tasks, and shipping them straight to prod!
Mostly digging through the Postgres manual and building a distributed database, from scratch, using first principles (and as many Wikipedia articles as you need). Everything is written in Rust, so that makes it easier, honestly. Type-safe, “fearless concurrency”, and an awesome open source ecosystem where you can find pretty much everything we’ll need.
Time split is going to be something like 30% bug fixes, 60% new features, 10% infra glue. Use as much (or as little) AI as you want. I use Claude a lot, and it’s been pretty helpful overall, especially for debugging what is now over 100k lines of code. 🚀🚀🚀
I did this at Instacart back in 2020. Grew the company 5x in a few months, just by making Postgres do things it wasn’t built for. My key insight: if you make the database you already have just work, people love it. Nobody wants to migrate to Dynamo, seriously.
I’ve been coding in Rust for close to 10 years, even wrote my own web framework for fun. PgDog started as “something to do” while I was unemployed; one year later, it’s mission-critical infra for some pretty serious people, so that’s cool.
Postgres is the most popular database period, but running it in production is hard. We're building scaling tools, like load balancing and sharding, to make Postgres the one database you'll ever need.