← All speakers

Bio, Work & Ideas

Michael Yuan

Michael Yuan is the founder and maintainer of WasmEdge, a WebAssembly runtime for cloud-native and edge applications, and a co-founder of Second State and Gaia Network. He builds open-source infrastructure that combines portable execution, locally hosted models, and compiler feedback to make AI-generated software more dependable.

Trained as an astrophysicist, Yuan worked at JBoss and subsequently Red Hat, later serving as a principal investigator on National Institutes of Health-funded digital-health projects. He also wrote six software-engineering books spanning mobile Java, enterprise applications, smartphones, and blockchain.

At Second State, which he co-founded and previously led as chief executive, Yuan expanded from blockchain execution environments into WebAssembly infrastructure. He founded WasmEdge to support lightweight, sandboxed applications across cloud, edge, and decentralized systems, and co-authored research on integrating WebAssembly with edge orchestration.

His subsequent projects extend that foundation into AI. LlamaEdge uses WasmEdge and Rust-based components for locally hosted AI inference and OpenAI-compatible services; Gaia Network applies related infrastructure to distributed models and agent-accessible applications.

  • Compiler-verified coding agents: Yuan argues that Rust’s strict typing and detailed diagnostics, difficult for many human beginners, give coding agents actionable feedback: generate a program, compile it, inspect failures, and revise it. Successful compilation strengthens confidence in generated code without proving complete correctness.
  • Agent-accessible software infrastructure: Yuan treats AI systems as a new audience for software interfaces. His work on OpenMCP.app translates existing APIs into agent-usable services, while his longer-term vision pairs those tools with generated, compiled, and repaired task-specific programs.

Read the topics behind these talks

1 conference talk

Key ideas

Scroll to read ↓

RustCoder turns Rust’s demanding compiler into feedback for code generation, combining retrieval, model-specific prompts, and repair tools that work through both IDEs and agent interfaces.

  • What changes when the programmer is a machine?
    0:00 ↗
  • The compiler becomes a feedback channel
    3:41 ↗
  • From a Rust learning corpus to a running program
    6:43 ↗
  • Giving Cursor a compiler-and-repair tool
    11:36 ↗
  • A broken Hello World makes the exchange visible
    14:33 ↗
  • Specialization lives in the whole repair stack
    16:43 ↗
  • Runtime, retrieval, and the MCP boundary
    19:51 ↗
  • Beyond the IDE: software consumed by models
    21:16 ↗
  • Running RustCoder and exchanging projects
    24:56 ↗
  • The remaining work is concrete Rust knowledge
    28:13 ↗

References