During March 2025, contributed to the nearai/nearai repository by implementing image-to-image generation for the Fireworks AI Service. This feature allows users to upload an initial image and adjust generation strength, supporting nuanced image edits within the platform. The solution involved decoding base64-encoded image inputs and conditionally routing requests to the appropriate Fireworks API endpoint, ensuring seamless integration with existing backend workflows. Leveraging Python for backend development and API integration, the work expanded the service’s image generation capabilities and enabled more flexible creative workflows for users. The implementation focused on robust preprocessing and end-to-end support for advanced image manipulation tasks.