Building Serverless Python AI skills as WASM components
summary
Claims that combining WebAssembly component model with serverless principles lets teams ship Python AI skills as WASM components that the runtime runs and secures. Implements a Python SDK, uses componentize-py / Bytecode Alliance tooling and WASI Python runtime, exposes a Cognitive System Interface (CSI) for inference, vector DB, streaming and tracing, and publishes artifacts to an OCI registry. Runtime (Faria kernel) runs these skills in production for about a year; components are deliberately constrained (no outbound HTTP) for safety and observability.
AI-written summary · about
chapters (8)
- 0:00Serverless AI with WASM
- 1:35AI in production challenges
- 7:11WebAssembly component model
- 12:47Python AI skills in WASM
- 21:00Packaging WASM skills
- 22:24CSI interface constraints
- 25:13Next steps with agents
- 29:58Q&A on MCP
related talks
31m"Serverless AI Inferencing Using Python and WebAssembly"
01h 04mDevelop, Build & Deploy Lightweight AI Services with WebAssembly
31mBuild a Serverless WebAssembly App in Python using Spin
43mBuilding Serverless AI Workflows with Wasm+Rust
08mKeynote: How WebAssembly can Power the New Wave of Serverless AI
32mFrom Cloud to Edge computing - Unleashing the power of WebAssembly at the edge
41mWebAssembly Based AI as a Service
31mRunning AI workloads with WASM - is it production ready? [Panel]