Nicholas Williams

Built a dbt-based semantic layer across Snowflake and Omni, documenting model definitions, lineage, and business meaning. Built or materially improved 50+ dbt models across Finance, pricing, Sales, claims, and Operations, moving repeated dashboard logic into reusable datasets. Designed governed self-service analytics agents used by 40 employees across five departments, reducing inbound data-team requests 22%. Created REMI's first financial snapshots and showed that mutable records were causing historical reporting drift.

Maintained and improved a transaction data engine processing about 130M transactions per day. Led a transaction-location initiative that raised physical-location accuracy from 83% to 91%. Used Python automation to reduce analyst dependence on Engineering.

Recreated production environments and maintained automated tests. Documented defects and gave developers usability and performance recommendations.