Or Avital developed AI-powered data validation features for the elementary-data/dbt-data-reliability repository, focusing on scalable data quality checks across Snowflake, Databricks, and BigQuery. He introduced a macro that leverages LLM integration to validate both structured and unstructured data against user-defined expectations, reducing manual validation efforts. Using SQL, Jinja, and dbt, he standardized parameter handling, improved naming consistency, and removed legacy code to streamline maintainability. Or also updated documentation in the elementary-data/elementary repository, clarifying feature capabilities and business alignment. His work demonstrated depth in AI/ML integration, data engineering, and technical writing, delivering robust, multi-cloud data validation solutions.