In September 2025, K. Markert developed the GoogleMapsGroundingTool for the Shubhamsaboo/adk-python repository, enabling Gemini model queries to be grounded with Google Maps data. The work focused on robust API integration and model grounding, ensuring compatibility with the VertexAI Gemini API and implementing explicit error handling for unsupported models or versions. Markert incorporated internal tool gating to maintain stability and controlled access, preparing the codebase for future enhancements. Using Python, the solution addressed the need for reliable, map-based grounding in Gemini 2, reflecting a focused engineering effort on tool development and forward compatibility within a production machine learning environment.