AI for Research: NemoClaw Lab Workshop

Friday 2 October · INOX

Each session lasts 90 minutes · Run four times: 9:00, 11:00, 13:30 and 15:30

A chatbot answers one prompt. An agent runs a loop: plan, call a tool, read the result, decide what to do next, repeat. The same task goes from a few thousand tokens to hundreds of thousands, takes minutes rather than seconds, and raises questions a chatbot never did — what it costs, where your data goes, and what the agent is allowed to touch.

This hands-on lab lets you find out by running one yourself. Working in a pair on a Dell Pro Max with GB10, you stand up the NVIDIA NemoClaw runtime and see what each part of the stack does: vLLM serving a model locally on the GPU, OpenClaw running the agent loop, NVIDIA OpenShell sandboxing it. You then give an agent tools, memory and a real workflow, run it, and watch tokens, cost and latency in real time — before connecting the local agent to data-centre AI services and seeing how the two fit together.

This is a technical session. You should be comfortable in a Linux terminal, know roughly what Docker does, have used an LLM through an API or a coding agent, and be able to read a short Python script and change it. You won't be building an application from scratch. Hardware, operating system, model and lab environment are all set up before you arrive.

Places are limited and you'll be sharing a machine with one other participant, so please only book if you can stay for the full 90 minutes.