Galaxy is a lab building data infrastructure for the age of AI.

Every company deploying AI hits the same wall. The models are ready. The data is not. Context is scattered across dozens of systems, nobody agrees on what a customer or an account actually is, and the tools meant to fix this have grown more expensive and more locked in every year.

We build the layer underneath. Systems that move data reliably, model it into something with meaning, and give both people and agents one grounded source of truth. Replication, ontologies, knowledge graphs, and the plumbing that holds it together at scale.

We are building these in the open, one component at a time.

The best infrastructure in the world is open. Engineers adopt what they can read, run, and trust, and what they trust becomes what everything else gets built on. So we publish our work, we publish our benchmarks, and we ship software a normal engineering team can deploy and operate inside its own environment, with no vendor holding the keys.

That matters more as AI gets deeper into how companies work. An agent is only as good as the ground it stands on. Owning that ground, the data, the pipes that move it, and the meaning layered on top, is what makes AI dependable at the scale a business actually runs at.

Alongside the open source work we operate as a forward deployed team, building and running data systems inside the companies we partner with. It is how we learn what is actually broken, and it is what we productize next.

Data engineering is becoming context engineering. We are building the infrastructure it runs on.