Migrating SQL workloads between different platforms like Databricks SQL and BigQuery is a common challenge due to variations in syntax and functions. This blog post details the process of translating Databricks SQL queries to BigQuery SQL. The author aimed to improve efficiency and reduce costs by moving analytics workloads from Databricks to BigQuery. Manual translation proved to be time-consuming and prone to errors. Google Gemini was employed as an AI assistant to bridge the gap between the two SQL dialects. The process involved creating a function mapping guide and using few-shot examples to train Gemini. A Retrieval-Augmented Generation (RAG) layer was implemented using Vertex AI to provide Gemini with contextual information, enhancing translation accuracy. The architecture included storing source SQL in Google Cloud Storage and integrating with the Gemini API. A validation layer with BigQuery dry runs was used to check for syntax issues. Key takeaways emphasized the effectiveness of combining RAG with Gemini, the importance of a comprehensive function mapping guide, and the necessity of thorough validation. This approach streamlines SQL migrations, making the process faster and more reliable.