Minh Duy Tran
Founder & Lead Multi-Agent Manga Systems Architect · Ho Chi Minh City & Tokyo Research Node · 9+ years experience
Lead architect of TextToManga's multi-agent storyboard director, persistent Character Bible conditioning pipeline, and bilingual CJK/Latin comic typesetting engine.
Background & Sequential Art Focus
Minh Duy Tran leads core architecture at TextToManga, designing the multi-agent pipeline that bridges long-form prose manuscripts (light novels, webnovels, and screenplays) with deterministic multi-panel sequential art.
Prior to launching TextToManga Studio, they engineered high-throughput computer vision and document layout systems, focusing on solving cross-panel character identity drift and automated CJK vertical (Tategaki) speech balloon geometry.
Verification Methodology: Every benchmark and workflow guide authored by Minh Duy Tran uses reproducible 3,000-word prose chapters tested across identical hardware and rubric criteria.
Core Research Domains
| Specialty Track | Applied Focus |
|---|---|
| Domain #1 | Multi-agent LLM storyboard decomposition |
| Domain #2 | Sequential character identity conditioning |
| Domain #3 | Japanese manga nemu (ネーム) pacing |
| Domain #4 | Vertical webtoon production pipelines |
| Domain #5 | Automated comic speech bubble typesetting |