2025

Established an AI-integrated production workflow, overseeing the coordination of art direction, 3D modeling, and post-production teams. Provided technical guidance in leveraging Kling, Veo3, Wan2.5, Trip 3D, Nano Banana, Midjourney, and Magnific AI for generating art design components, 2D animation sequences, and 3D assets for the lighting show. The workflow achieved an AI participation rate of over 90%, representing a major advancement in content automation and creative efficiency.

2019

In response to the client’s demand for an advanced delivery timeline, an automated video transcoding pipeline was designed and implemented. The system converts unified visual productions into 954 separate LED playback files, corresponding to each building’s control system.

Through the integration of digital imaging concepts and programmatic development principles, a Unity- and FFmpeg-based system was jointly developed with software engineers. By referencing satellite map data, the system realized semi-automated spatial output and precise control of 954 individual building visuals, enabling the entire urban landscape to respond rhythmically to music.

In addition, I was responsible for the final compositing, editing, and delivery of the finished visuals.

2024

Became one of the first creators in China to deploy and complete an entire projection show production independently using the ComfyUI visual workflow platform, integrating AI-assisted compositing and generative content design for large-scale architectural projection.