Generative AI

This space explores the integration of generative AI across social media, data visualization, and visual effects pipelines. My practice focuses on developing hybrid workflows: building custom pipelines that maintain strict character consistency while preserving a manual, human touch, and utilizing AI tools as a final-mile renderer for complex cinematic assets and shot breakdowns.

Ultimately, I am interested in blending data-driven visualization with AI, using human-driven data structures to guarantee absolute structural accuracy while testing the potential of AI to deliver the final aesthetic.

Ggamnyang │YouTube Channel

A short-form social media project focused on building custom generative AI workflows around a central character, a distinct black kitten named Ggamnyang.

The core technical challenge involved bypassing the random variance of AI generation to maintain strict character consistency across different scenes. Manual editing and post-production were integrated throughout the pipeline to give the final pieces a natural human touch and feel.

Tools Used: Nano Banana, Hailuo, Runway

Death Star Sequence

An experimental visual effects sequence integrating a traditional 3D layout pipeline with generative AI tools to deliver a high-fidelity cinematic shot.

The project focuses on using AI strictly as a final-mile renderer. By establishing the structural composition and camera movement manually in 3D, the generative AI pass was utilized to upscale textures, enhance atmospheric details, and blend the final elements together without losing hard-surface architectural control.

Tools Used: Houdini, Nano Banana, Kling