About Lyric
Lyric is building an AI-native, composable decision platform for the world’s most complex supply chains. Our software is used by enterprises including DHL, PepsiCo, Mondelez and Exxon. We were founded by the team behind Opex Analytics and have raised $80M.
The role
We’re hiring a hands-on Platform Engineer – Data at SDE-3 / Staff level to take a billion-row shared data platform from 10 to 100. This is platform ownership, not business ETL or analytics-pipeline delivery. You’ll design, build and operate the foundations that product and engineering teams depend on.
What you’ll own
• Open-lakehouse foundations across Iceberg, Delta Lake or Hudi: table lifecycle, partitioning and layout, compaction, schema evolution, catalog and query operations
• Batch and streaming ingestion, including CDC, data contracts, deduplication, checkpoints, replay and recovery
• Services, APIs and self-service primitives that make the platform usable without creating a central-team bottleneck
• Orchestration and workload execution across Kubernetes and tools such as Prefect, Argo or Airflow
• Metadata, lineage, governance, data quality and observability
• Production reliability: incidents, SLAs, freshness, cost/performance trade-offs and faster recovery
What we’re looking for
• Typically 7+ years of experience, with clear SDE-3/Staff-level scope even if your current title differs
• Hands-on, end-to-end ownership of a shared data platform in production—not only pipelines built on top of one
• Strong distributed-systems design and production coding in Python, Java, Go or Scala
• Meaningful depth across batch, streaming, ingestion/CDC, storage/query systems and platform operations
• Evidence that you improved reliability, performance, cost or developer self-service at scale
• An IC who can drive architecture and technical standards while staying close to code and operations
Strong adjacent signals
Open-source or deep production work with Iceberg/Hudi/Delta, Kafka/Flink/Debezium, Spark/Trino/Pinot, Kubernetes, data catalogs or observability systems.
Not the fit
BI/reporting, analytics engineering, warehouse modelling, business-facing ETL-only work, generic backend, managed-service assembly, narrow single-component ownership or a people-management-only role.
