At Cantar Analytics, I built 30+ dbt models in Snowflake for supply-chain data, implementing a star schema and 114 automated tests. The models give analysts and Power BI dashboards a trusted reporting layer.
I also built an LLM-based Text-to-SQL system over Snowflake. It achieved 80% query accuracy across 100 test questions, using schema-aware prompting and read-only SQL guardrails.
For my E-commerce Event-Driven Data Pipeline project, I automated ETL with JSON-triggered Databricks Workflows, reducing manual processing effort by 70%. I implemented SCD Type 2 history with Delta Lake MERGE operations and added PySpark schema validation and file archiving to prevent duplicate processing.
On my Airline ETL & Data Ingestion Pipeline project, I built an incremental pipeline handling 0.5 million daily flight records. I used AWS Step Functions and EventBridge to orchestrate ingestion, with SNS notifications for job failures and completions.
