We got tired of being the last ones to know when a pipeline broke.
Stratum started as a set of internal scripts we wrote to survive on-call rotations. When we realized every data team we knew had the same problem, we decided to turn it into a product.
Built from the
on-call rotation.
In 2022, Amir was the data platform lead at a Series B company. Every Monday morning started with the same ritual: checking Slack to see which dashboard had broken over the weekend, and then spending two hours tracing the failure back through Airflow logs.
He and Jana — then a senior data engineer at Stripe — had both built internal tools to catch issues earlier. Theirs worked. But they were duct tape: brittle, undocumented, and the first thing to break when someone changed a schema.
They started building Stratum in the evenings. The first version was a Slack bot that diffed your dbt schemas between runs. Teams loved it. The second version added pipeline versioning. The third version added environment promotion gates.
Today, Stratum is the tool they wished had existed. If you have ever been paged at 2am because a column got renamed upstream, this was built for you.
How we think about
building this.
Calm over chaos
Pipelines fail. The goal is to know about it before anyone else, handle it without panic, and understand exactly what changed. Our product is designed around that philosophy.
Boring reliability
We would rather be the tool you forget about because it always works than the tool you talk about because it has interesting features. Reliability is the feature.
Engineers first
Stratum is built for data engineers, not for dashboards or VPs. Every decision goes through the question: would a data engineer trust this?
Simple by default
You can get value in 10 minutes or 10 hours. We try not to hide depth, but we never require it upfront. Start simple, grow into more.
The people behind it.
Amir Moradi
Co-founder & CEO
Former data platform lead at Segment. Spent three years on-call for Airflow failures before deciding to fix the problem properly.
Jana Torres
Co-founder & CTO
Built data infrastructure at Stripe and dbt Labs. Believes most pipeline incidents are a monitoring problem, not an engineering one.
Sam Reeves
Head of Design
Previously at Linear and Vercel. Thinks tools for data engineers should be as well-crafted as tools for software engineers.
Kofi Osafo
Head of Customer Success
Analytics engineer turned customer advocate. Has personally onboarded over 80 data teams and knows every edge case by name.
We are hiring.
Remote-first. Looking for people who care about data reliability as much as we do.