Data Engineer

About the role

Auralis ingests subscription and billing events from a few hundred customers, each with their own idea of what a clean payload looks like. Your job is to turn that into a warehouse our product and our customers can trust without an asterisk.

The interesting part

Almost nothing arrives cleanly. Events land late, get replayed, carry the wrong timezone or reference an entity that has not been created yet. Designing models that stay correct through all of that, and that make a bad batch visible rather than silently averaged away, is most of the work.

The stack is Python, Airflow, dbt, Postgres and a columnar warehouse, with Spark for the two customers whose volume needs it. We deliberately keep it small enough that one engineer can hold the whole pipeline in their head.

What we look for

Three to six years in data engineering or a strong backend background with real SQL depth. You should be opinionated about modelling, comfortable with orchestration, and disciplined about testing data the way you would test code.