You'll own the data layer behind our analytics product and design of our analytics architecture and data model. Today that's a high-performance analytical store serving customer-facing queries; where we're headed is a tiered platform that separates our system of record, a low-latency serving tier, and a batch/ML tier. You'll keep the serving layer fast and reliable while designing and building that broader architecture, deciding where each workload lives based on measured latency, cost, and isolation.
This is a hands-on engineering role with strong architectural influence. You'll write code and also shape the multi-quarter direction of the analytics platform.
Find and fix database performance and scaling problems.
Own the schema and data model for our analytical data.
Design and build our analytics architecture: a tiered platform spanning our system of record, a low-latency serving tier, and a batch/ML tier.
Decide where each workload runs across those tiers, backing placement decisions with real benchmarks rather than vendor claims.
Build observability into the data layer so problems surface early.
Partner with the ingestion team to keep the read and write paths coherent as schemas evolve.
Deep, production experience operating a database or analytical store under concurrent, customer-facing load: you can reason from a query plan to a fix, not just from documentation.
Strong SQL and a real mental model of columnar / OLAP execution.
Solid backend engineering. Our stack is Java, but we care more that you can read and modify a real codebase than which tools you've used.
Data architecture experience: you've designed or materially shaped a multi-tier data platform and can defend those decisions with measurements.
Strongly preferred
Lakehouse experience: Databricks, Spark, Delta Lake / Iceberg, or comparable.
Experience designing schemas for wide, semi-structured event data (nested objects, maps, JSON).
Streaming ingestion experience (Kafka or similar).
Track record of a migration or major re-architecture, with the cost/latency reasoning to back the decisions.
ClickHouse experience.
Connection-pool and JDBC-level tuning experience.
Familiarity with feature-flagged rollouts of query-engine behavior changes.
