# Tae Kyung Kim

- **Event:** [Gemini 3 서울 해커톤](https://cerebralvalley.ai/e/gemini-3-seoul-hackathon)
- **When:** Sat, Feb 28 at 9:00 AM – 10:00 PM (GMT+9)
- **Where:** 서울 Seocho District, Olympic-daero, 2085-14 세빛둥둥섬
- **Team:** [Tae Kyung Kim](https://cerebralvalley.ai/u/aigement)
- **GitHub:** https://github.com/snutp/gemini-hackathon-seoul
- **Demo video:** http://211.192.160.149:8090/static/demo.mov (open in new link / tab)
- **Gallery:** https://cerebralvalley.ai/e/gemini-3-seoul-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/gemini-3-seoul-hackathon/hackathon/gallery/73

Link: http://211.192.160.149:3020/

The current mainstream process for designing LLM-generated outputs (PPTs and infographics) remains quite rudimentary.

  ChatGPT, Claude, and Gemini (despite their recent major improvements in SVG
  output), along with wrapper services such as Genspark, still tend to produce
  designs that are only at the level of undergraduate coursework. As a result,
  they are largely unusable in academic and business settings.

The biggest problem, however, is that generated images are essentially non-
  editable. To address this, I built a fully web-operable infographic image decomposer using visual grounding, text-font matching, and SVG reconstruction into editable
  components. 

  When this is connected to Nano Banana’s infographic generation pipeline, it
  enables a workflow that is fully automated yet still supports high-end post-
  generation polish.

---

Markdown version of https://cerebralvalley.ai/e/gemini-3-seoul-hackathon/hackathon/gallery/73. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt