Pranav Karmarkar enhanced the google/adk-java repository by adding support for image and media attachments in the MessageConverter, enabling visual inputs in AI requests. Using Java and the Spring Framework, he addressed a gap where media content was previously ignored during the ADK-to-Spring AI handoff, improving compatibility with vision-enabled models such as GPT-4o. His approach included updating the UserMessage construction to handle media attachments and developing targeted unit and integration tests to validate robust media processing. This work expanded backend AI workflow capabilities, reduced manual preprocessing, and allowed for more reliable and versatile interactions with vision-enabled AI models.