Hands-On Lab Workshop: Hermes Agent - The Agent That Remembers You
Event Overview
Most AI assistants forget everything the moment you close the tab. In this session, we'll show you an assistant that doesn't: the Nous Research Hermes Agent, powered by NVIDIA's Nemotron-3-Super-120B-A12B model in WWT's AI Proving Ground. Hermes keeps persistent state on disk, remembering who you are, how you like to work, and what you've taught it, across every session and across surfaces including a desktop app, a chat platform, and the terminal. Join us as we commission the agent as an infrastructure co-pilot, teach it engineering standards, explore built-in skills like architecture diagrams and implementation plans, and save a brand-new reusable skill, then watch it apply everything automatically in a live virtual environment from a completely different surface.
Featured Speakers
What to expect
- A live walkthrough of deploying the Hermes Agent from the command line and connecting it to an inference endpoint
- A look at wiring up the Mattermost gateway so the agent runs in the background across sessions
- A demonstration of persistent memory in action, including frozen-snapshot Markdown files and full-text session search
- A real-time example of teaching the agent engineering standards and watching it apply them without a reminder
- A demo of turning a multi-step workflow into a saved, reusable skill the agent can recall on demand
- Live Q&A with presenter Ray Skinner
Goals and Objectives
• Install and configure the agent, connecting it to an inference endpoint and the Mattermost gateway • Commission a working agent personality and confirm it persists across surfaces without reconfiguration • Explore persistent memory: frozen-snapshot Markdown files and full-text session search beyond the snapshot • Teach the agent infrastructure reporting standards and see it follow them in a brand-new session • Save a multi-step workflow as a reusable skill and verify it appears in the Skills and Tools pane • Work with the same agent, memory, and skills across the terminal, Mattermost, and the desktop app
Who should attend?
• Infrastructure and platform engineers curious about AI copilots that retain context across sessions • AI practitioners and engineers exploring agentic workflows and persistent-memory architectures • Anyone interested in applying custom engineering standards and reusable skills to an AI agent If you want an AI assistant that actually remembers your standards, this session is for you.