Unison Group
Data Architect - Banking
- Singapore
- Contract
- Posted
Skills
Job description
We are looking for a Data Architect with 10+ years in data, including 4+ years in banking and a hands-on data engineering background.
Must-have skills
Banking and domain
•Data platform design inside tier-1 or large regional banks.
•Banking data models (Teradata FSLDM or equivalent) across customer, account, payments and AML/CFT subject areas.
•Financial crime or regulatory data marts built to BCBS 239 standards for lineage, reconciliation and audit.
Data architecture
•Enterprise data warehouse, data lake and Lakehouse architecture; Data Mesh and Data Fabric concepts.
•Data modelling: dimensional, 3NF and Data Vault; semantic layers and data products.
•Historical data design: snapshots, SCD Type 2, bi-temporal models and large multi-grain backfills.
•Data-as-a-service patterns across batch and API delivery, with lineage, observability and SLA monitoring.
•Metadata, business glossary, lineage and data quality frameworks.
Data engineering (hands-on)
•ETL/ELT development, CDC, historical data processing, metadata-driven automation frameworks and performance optimisation.
•Teradata and Hadoop/Cloudera at scale: warehouses of 100 TB+ and tables of billions of rows.
•Spark, Python, R, SQL, Hive, Impala, HBase, Pig and Hue; Apache Iceberg and Trino.
•Streaming and integration: Kafka, Spark Streaming, Informatica (including IDL) and Talend.
•Databases and data federation: advanced SQL, NoSQL, Denodo and Snowflake.
•Cloud and DevOps: Kubernetes, Airflow, GitOps and CI/CD tools.
Delivery and communication
•Running design authority forums and reviewing engineers’ designs and code.
•Writing design specifications that pass architecture, audit and regulatory review.
Good-to-have skills
•Databricks, Snowflake and Lakehouse platforms; data services on Azure, Google Cloud, Huawei Cloud or Alibaba Cloud.
•BI and analytics platforms: Qlik Sense, Microsoft Power BI and Azure analytics services.
Certifications (preferred): TOGAF; Teradata Vantage; Cloudera; DAMA CDMP; Databricks or cloud architect