Michael Yuan (@juntao) 's Twitter Profile
Michael Yuan

@juntao

Run LLMs and multimodal AI models on your own device and / or embedded in your own agent apps! github.com/LlamaEdge/Llam…

ID: 14057709

linkhttp://michaelyuan.com calendar_today28-02-2008 22:41:26

1,1K Tweet

2,2K Takipçi

901 Takip Edilen

Michael Yuan (@juntao) 's Twitter Profile Photo

Running the Gemma-3 VLM on LlamaEdge. The 4b model is small enough for most edge devices and yet it is capable of complex visual reasoning. This feature will come to your Gaia node Gaia 🌱 soon! WasmEdge wasmedge provides a lightweight and cross-platform runtime for

Michael Yuan (@juntao) 's Twitter Profile Photo

Is AI replacing doctors? Especially radiologists? The MedGemma is a very small LLM that could easily run on a mobile device. It is trained on medical images and knowledge. I tried it. It is quite remarkable. Watch the video 👇 You can run MedGemma and the general purpose Gemma

Antoine (@antoinegrondin) 's Twitter Profile Photo

the GenAI stack presented by Michael Yuan who works on WasmEdge. his argument: AI is making workloads less portable, bucking the recent trend toward interoperability. sidecars are inefficient. HTTP is stateless but AI is stateful. Wasm is the answer? succesful use-cases. OpenAI

the GenAI stack presented by <a href="/juntao/">Michael Yuan</a> who works on WasmEdge.

his argument: AI is making workloads less portable, bucking the recent trend toward interoperability.

sidecars are inefficient. HTTP is stateless but AI is stateful.

Wasm is the answer? succesful use-cases. OpenAI
Gaia 🌱 (@gaianet_ai) 's Twitter Profile Photo

The first phone where your AI never leaves your device. No cloud processing. No data harvesting. Complete AI sovereignty. Built on Galaxy S25 Edge hardware. Earn rewards through the Gaia network. 1,000 units now available. Additional releases planned. mobile.gaianet.ai

Michael Yuan (@juntao) 's Twitter Profile Photo

This AI device is already used in classrooms from top universities (including #1 ranked CS program by U.S. News & World Report) to elite high schools. Everything is open source. Reserve yours now.

Michael Yuan (@juntao) 's Twitter Profile Photo

The most accomplished logician in modern history, Kurt Gödel, almost failed his civics test when he interviewed for US citizenship. He claimed to have identified a logical flaw in the constitution that would allow the US to become a dictatorship. Einstein had to stop him and bail

Michael Yuan (@juntao) 's Twitter Profile Photo

At the embedded AI workshop in the Rust Global Conf, students and developers loved EchoKit - an open-source hardware and software toolkit written in Rust Language - to build and learn about real-world AI agents. 🚀 echokit.dev

At the embedded AI workshop in the Rust Global Conf, students and developers loved EchoKit - an open-source hardware and software toolkit written in <a href="/rustlang/">Rust Language</a> - to build and learn about real-world AI agents. 

🚀 echokit.dev
Michael Yuan (@juntao) 's Twitter Profile Photo

EchoKit now supports ElevenLabs as a TTS service. Listen to it answering my question about ElevenLabs Is it correct? 😂 Get it: echokit.dev/echokit_diy.ht… Quick start: echokit.dev/docs/quick-sta… How to: echokit.dev/docs/dev/echok…

Michael Yuan (@juntao) 's Twitter Profile Photo

With Docker support, E2B is no longer just a sandbox environment for running LLM generated code (code interpreter). It now also runs LLM generated tool calls through the MCP protocol. Those MCP servers are already Dockerized and they can now be started in the same E2B

freeCodeCamp.org (@freecodecamp) 's Twitter Profile Photo

Voice apps have pretty strict requirements for latency, privacy, and customization - understandably. So it's often hard to find an AI voice tool that really fits your needs. In this tutorial, Michael Yuan helps you build your own voice AI agent using open source tools.

Voice apps have pretty strict requirements for latency, privacy, and customization - understandably.

So it's often hard to find an AI voice tool that really fits your needs.

In this tutorial, <a href="/juntao/">Michael Yuan</a> helps you build your own voice AI agent using open source tools.