top of page

takt op. Symphony

takt-op-feature-750x392.jpg

As a Mobile App Development Intern on takt op. Symphony with the Game Tech Center at DeNA, I focused on URP volumetric lighting, AIGC research, and cross-device QA testing to support visual fidelity and performance optimization across mobile platforms.

CONTRIBUTIONS

  • Developed a custom volumetric lighting effect using Unity’s Universal Render Pipeline

  • Explored AIGC workflows with Stable Diffusion and Midjourney, deploying models on AWS and fine-tuning LoRA models with 2 other interns

  • Conducted performance benchmarking on multiple Android devices

Tools

Unity, C#, HLSL

Technical Art

One of my key contributions during the development of takt op. Symphony was creating a custom volumetric lighting effect within Unity using the Universal Render Pipeline (URP). The goal was to enrich the visual depth and mood of key scenes while maintaining performance across mobile devices.

​

​

Responsibilities:​

  • Designed and implemented a custom volumetric lighting system leveraging Unity’s URP framework and Shader Graph, simulating light scattering and atmospheric fog

  • Collaborated with the art and tech teams to tune lighting parameters that enhanced emotional tone without compromising frame rate

  • Ensured mobile compatibility by optimizing shader performance and testing across a variety of Android devices using Unity’s UPR tools

​

AIGC

In this project, I conducted an in-depth investigation into AIGC workflows, focusing on . My exploration compared Stable Diffusion and Midjourney in terms of control, output quality, and adaptability for game production.

​

  • Deployed Stable Diffusion models on AWS with custom configurations to enable on-demand image generation with local control Fine-tuned multiple LoRA models to specialize image outputs for in-game use cases such as character design and stylistic consistency

  • ​Compared performance and usability of Midjourney vs. Stable Diffusion

  • Built internal tools to streamline AIGC-assisted workflows and collaborated with team members on creative prompt engineering

​

​This research laid the groundwork for scalable AIGC pipelines that support iterative concept development, enabling faster experimentation while maintaining artistic direction.

bottom of page