← All speakers

Bio, Work & Ideas

Christopher Chedeau

On this page

Christopher Chedeau, known as Vjeux, created Excalidraw, CSS-in-JS, the Yoga layout engine, and React Conf, and co-created React Native and Prettier. His software has shaped how developers build interfaces, format code, collaborate visually, and incorporate AI without losing control of their work.

From gaming communities to developer infrastructure

Educated at EPITA in France, Chedeau worked on Curse.com and MMO-Champion and created SC2Mapster before joining Facebook. There he contributed to photo tagging and image-serving infrastructure, helped open-source React, organized the first two React Conf events, and started Create React App.

On the original React Native team, he helped define component interfaces and animation tools for native mobile applications. CSS-in-JS emerged partly from a practical decision: expressing styles in JavaScript avoided building a traditional CSS parser and selector engine. He also developed the first version of Yoga to bring consistent layout behavior to native interfaces.

Chedeau subsequently helped turn James Long’s experimental JavaScript formatter into Prettier. His history of the project describes the exacting work behind trustworthy formatting: snapshot tests, repeated-formatting checks, editor integration, continuous integration, and gradual adoption. Automating stylistic decisions also reduced recurring arguments over indentation, whitespace, and semicolons. He has additionally highlighted the difficulty of funding maintainers after free software becomes essential infrastructure.

Excalidraw and practical AI

On January 1, 2020, Chedeau prototyped Excalidraw using Rough.js after the drawing tool Zwibbler became unavailable. Initially intended for blog illustrations, the open-source project quickly became a collaborative whiteboard as pandemic-era teams moved online. Its expandable canvas, typed text, reusable shapes, object-level editing, and copy-and-paste exploited capabilities unavailable on physical whiteboards.

His approach to AI-assisted whiteboarding emphasizes three concrete principles:

  • Structured intermediate representations: When direct diagram generation proved unreliable, Excalidraw used Mermaid syntax as an intermediate language, then converted structured descriptions into native elements through conventional software.
  • Editable AI-assisted diagrams: Generated diagrams remain working documents whose objects, colors, and layouts people can modify, preserving human control over the finished artifact.
  • Workflow-specific automation: Context-aware document names and useful illustrations address actual whiteboarding tasks; impressive image-generation demonstrations do not automatically belong inside a diagramming product.

In 2026, Chedeau extended these ideas to an AI-assisted TypeScript-to-Rust port of Pokémon Showdown. Code that merely compiled was insufficient: preserving behavior required deterministic scaffolding, smaller files, persistent instructions, reproducible end-to-end tests, and sustained human architectural judgment.

Read the topics behind these talks

1 conference talk

References