Staff Engineer, Tracing.

About the team

Tracing is five engineers across Rotterdam, Lisbon and Kraków. We own the pipeline that records every block of every run, the store that keeps it and the trace view people open when something breaks. Last month that was 2.1 billion runs.

A trace opens in 162 ms at p95 and we have not lost a span in twelve months. Next we want 30-day replay on Team and trace export into customers' own warehouses.

What you will do

• Set the technical direction for ingest, storage and query across 12 regional clusters.

• Extend replay from seven to 30 days of real events without raising the bill.

• Design trace export to Snowflake and S3 so customers can keep spans as long as auditors ask.

• Join the on-call rotation one week in six, with a paid day off after each shift.

• Review design docs across Engineering and mentor the engineers who write them.

What you bring

• Eight or more years building data systems that run in production at scale.

• Deep experience with a columnar store such as ClickHouse, Druid or BigQuery.

• Strong Go or TypeScript, and a habit of measuring before you change things.

• Clear written English. Most of our decisions start as a document.

Nice to have

• You have built observability or tracing tools that other engineers relied on.

• You have run a storage system through a schema migration with no downtime.

• You have worked under data residency rules in more than one region.

Our stack

Go and TypeScript, Kafka for ingest and ClickHouse for storage, deployed on Kubernetes in 12 regions. We use GitHub, Linear and Notion, and we deploy every weekday.