"Serverless AI Inferencing Using Python and WebAssembly"
summary
Demonstrates using Spin and the Python SDK (componenti_py) to compile Python handlers into WebAssembly components, build with spin build and deploy to a serverless runtime, and run LLM calls (example: Llama 2) and a sentiment-analysis endpoint. Shows a MobileNet benchmark: Linux container TensorFlow ~200+ ms versus wasmEdge <50 ms (5–10× speedup). Live demo builds app.wasm, deploys to a cloud workers-style host, invokes GPU-backed inference per-request, and returns working responses.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 4:35WebAssembly benefits
- 10:13Spin framework
- 11:58Cold starts with WASM
- 19:26Python serverless AI app
- 22:16Live demo
- 30:27Q&A
related talks
25mBuilding Serverless AI Apps with Spin and WebAssembly
31mBuild a Serverless WebAssembly App in Python using Spin
01h 59mCreate a First AI Program with TypeScript and Serverless WebAssembly
08mKeynote: How WebAssembly can Power the New Wave of Serverless AI
01h 34mTutorial: Getting Started with Serverless WebAssembly and Spin
35mBuilding Serverless Python AI skills as WASM components
01h 04mDevelop, Build & Deploy Lightweight AI Services with WebAssembly- 01h 34mTutorial: Getting Started with Serverless WebAssembly and Spin