Worked on the archestra-ai/archestra repository to implement automatic synchronization of OpenRouter embedding models, enabling seamless integration of embedding models into the model catalog. Leveraged TypeScript for backend development, focusing on robust API integration and schema validation using Zod to ensure data integrity during bearer-authenticated fetches. Enhanced the model catalog pipeline by expanding fetch paths and inferring embedding dimensions for known model IDs, which reduced manual synchronization efforts and improved lifecycle reliability. Updated provider documentation and introduced targeted tests to validate model fetching, embedding dimension inference, and schema validation, laying the groundwork for reliable embedding model delivery and faster onboarding.