Developed and delivered a feature for the vercel/ai repository that enables embedding of remote video, audio, and PDF content stored in Google Cloud Storage. This work involved extending the Google embedding model schema with fileData support, allowing users to reference external assets through fileUri and mimeType fields. The implementation was completed using TypeScript and involved both API and full stack development, with comprehensive end-to-end and integration testing to ensure reliability. Collaboration with other contributors ensured timely delivery and maintained CI health, broadening the SDK’s multimodal capabilities and streamlining enterprise workflows that require referencing external media assets.