You’ve built and operated backend systems that real customers depend on. You can work through questions about data, APIs, performance, and security, and explain the consequences of your decisions.
Experience integrating AI technologies like OpenAI and Pinecone into production systems to build innovative, data-driven solutions. Experience with vector databases, transformers, NLP and LLMs is a plus.
Our stack includes TypeScript, NestJS, GraphQL, Postgres, and Prisma. We also work with OpenAI and Turbopuffer. Experience with these tools is useful; we’re especially interested in what you’ve built with them and the decisions you made along the way.
What we value
“How fast can we ship this?” comes up a lot at Pandektes. So does “we should take our time with this one.”
Knowing which question to ask is a big part of being an engineer here.
Knowing when it’s good enough
We like getting things into customers’ hands quickly. That means making trade-offs, questioning scope, and accepting when something is good enough for now. There’s usually another approach to explore or another prompt to try. At some point, you have to ship.
We use AI throughout our development process. AI code reviews, feature flags, and remote coding agents running end-to-end tests help us move quickly with confidence. Deciding what to build, what to check, and when it’s ready still takes judgment.
We value people who can find a smaller version worth shipping, explain the trade-offs, and follow through once customers start using it.
Knowing when to slow down
Some decisions deserve more time. Getting them right makes everything that follows easier. Getting them wrong means paying for the same mistake repeatedly.
Our default is to move quickly, so slowing down takes discipline. Sometimes it means pushing back when the rest of us want to get something out. We value someone who can explain why the extra work matters, challenge an approach, and help us decide where to spend our time.
You’ll have room to make those calls here, and teammates who will question them. We expect you to do the same for us.
Make time to experiment
We enjoy trying things before we know whether they’ll work. A new tool, a different way to use an agent, an idea that might turn out to be useless. Finding out is part of the work.
Once a month, we get together to demo our setups, show what we’re experimenting with, and share what we’ve learned. Sometimes we invite people from outside the team to show us how they work. Those sessions have changed how we build software.
We value curiosity and the generosity to bring other people along. If you find something useful, show us. If an experiment fails, we’d like to hear about that too.
We try to watch this video once a month to keep this value top of mind.
