Describe your experience building ETL/ELT pipelines. How do you handle data quality issues, retries, and schema evolution?