Leveraging Wasm for Portable AI Inference Across GPUs, CPUs, OS & Cloud-Nativ...
summary
Demonstrates a practical stack that lets developers build on CPU, run on GPU without recompilation by shipping models as a Wasm bundle that the runtime auto‑offloads to available accelerators; uses Llama.cpp / llama‑age integrations, a Wasm runtime (WasmEdge-style) and a W3C‑compatible inference API. Shows working demos (virtual pastor, coding assistant), runs same Wasm binary on M2 Mac and NVIDIA servers, removes heavy Python deps, and provides a secure sandboxed deployment path.
AI-written summary · about
chapters (12)
- 0:00Portable AI inference with Wasm
- 2:28AI knowledge-sharing services demo
- 5:47AI coding assistant for R demo
- 8:53AI app development in minutes demo
- 11:00Deploying AI across environments
- 14:37Why Python and traditional approaches fall short
- 16:29Hardware issues for AI inference
- 17:57Wasm as cross-platform solution
- 19:25Wasm for AI model deployment
- 22:19Cross-platform AI inference demo
- 23:53AI integration with Wasm APIs
- 24:36Wasm security and management
related talks
33mEfficient and Cross-Platform LLM Inference in the Heterogenous Cloud
55mFast And Portable LLM Inference With WebAssembly And Rust
37mGetting Started with AI and WebAssembly
27mWasm Is Becoming the Runtime for LLMs
31mRunning AI workloads with WASM - is it production ready? [Panel]
48mWorkshop: Efficient and Portable AI / LLM Inference on the Edge Cloud
01h 04mDevelop, Build & Deploy Lightweight AI Services with WebAssembly
40mWebAssembly for AI Infra : A Lightweight, Fast, and Secure Alternative Approach