Senior Data Architect

The core responsibilities for the job include the following:

Data lake architecture:

• Own the lakehouse design: storage layout, open table format (Iceberg, Delta Lake, or Hudi), catalog, and the medallion/zone model from raw ingest through to curated, query-ready datasets.
• Define partitioning, clustering, and compaction strategies tuned for time-series access patterns, long time-range scans over narrow tag selections, and recent-window queries.
• Solve the small-files problem that streaming SCADA ingestion inevitably creates, and set file sizing, sort order, and maintenance job standards.
• Design schema evolution and versioning so tag additions, renames, and vendor upgrades don't break downstream consumers.
• Build rollup, downsampling, and retention tiers: raw at full fidelity, aggregates for dashboards, archival for regulatory, and RAMS analysis.
• Choose and own the query layer (Trino/Presto, Spark SQL, DuckDB, or warehouse engine) and tune it for the workloads that actually matter.

Database engineering:

• Own the relational and time-series databases behind the platform schema design, indexing, query tuning, connection, and resource management.
• Capacity planning, HA/replication, backup, and tested restore procedures, upgrade paths.
• Define how the lake and the operational stores coexist: what stays hot in a time-series DB, what lands in the lake, and how the two reconcile against source historians.

Governance and operations:

• Data cataloguing, lineage, access control, and audit, including how OT-sourced data is classified and who can reach it.
• Cost and storage growth management [cloud spend / on-prem capacity, as applicable].
• Set standards, review designs, and mentor the data engineering team.
• Work with OT and network teams on data movement across segmented zones (Purdue model, DMZ, unidirectional gateways) without weakening the security boundary.

Requirements:

• 5+ years in data engineering, database engineering, or data architecture, with 3+ years specifically on data lake or lakehouse platforms in production.
• Deep expertise in at least one open table format (Apache Iceberg, Delta Lake, or Apache Hudi), including maintenance operations, snapshot management, and failure modes.
• Strong object storage fundamentals (S3, ADLS, GCS, or on-prem MinIO/Ceph layout, consistency, lifecycle policies, cost, or capacity implications).
• Distributed query engines: Trino/Presto, Spark, or equivalent, with real tuning experience.
• Expert SQL and strong RDBMS administration (PostgreSQL, Oracle, SQL Server, or similar): you can read a query plan and fix what's wrong.
• Time-series data at scale: TimescaleDB, InfluxDB, ClickHouse, or a historian, and a clear view of where each fits.
• Metadata and catalog systems (Hive Metastore, AWS Glue, Unity Catalog, Nessie, Polaris, or similar).
• Python and a JVM language, or equivalent depth for platform-level work.
• Infrastructure literacy: Linux, containers, Kubernetes, IaC, CI/CD.
• Track record of setting technical standards other engineers build on.

Strongly preferred:

• Industrial or OT data experience historians (AVEVA/OSIsoft PI, Wonderware, GE Proficy, Siemens WinCC, Ignition), tag namespaces, and asset hierarchies.
• Railway or metro domain experience in traction power, signalling and interlocking, tunnel ventilation, ECS/BMS, rolling stock condition monitoring, AFC, and depot systems.
• Air-gapped or on-premise deployments and the constraints that come with them, OT security standards (IEC 62443), and segmented network architectures.
• ISA-95 or equivalent asset information modelling, dbt, data quality frameworks (Great Expectations, Soda), and lineage tooling (OpenLineage, DataHub, Amundsen).
• Rail assurance standards (EN 50126/50128/50129) or other regulated engineering environments.