Abhigelot Gelot upgraded the Gemini Embedding Service in the emcie-co/parlant repository by migrating from text-embedding-004 to gemini-embedding-001, focusing on improving embedding quality and efficiency. Using Python, he refactored the service to update model names and configurations, reducing the maximum token count from 8000 to 2048 and increasing embedding dimensions from 768 to 3072. This configuration management work established a maintainable, upgrade-friendly architecture, enabling smoother future model transitions. The project emphasized API integration and maintainability, with no major bug fixes during the period, reflecting a targeted approach to feature delivery and long-term service reliability.