The Challenge
Where things stood before
Snowflake credit spend was up 40% YoY with queries missing SLAs. Micro-partitions were poorly clustered and ETL patterns forced full-scans on multi-billion-row tables.
Our Approach
How we engineered the outcome
- Profiled the top-100 warehouse queries via Query History and identified pruning gaps.
- Introduced multi-column cluster keys, materialized views and search-optimization service where cost-justified.
- Migrated legacy stored-procedure ETL to dbt models with tests, docs and CI/CD.
- Standardized parameterized ADF pipeline templates reusable across use cases.
The Outcomes
Measurable business impact
35%
Lower Snowflake credits
~50%
Faster query performance
30%
Reduced pipeline downtime
1,500+
Automated header files
