Interest in Project #43 Edge Orchestrator Agentic AI Integration

[TravisHaa]

Hello @scottmbaker @effndc, I hope this finds you well! I'm Travis Ha, a 3rd year computer science and engineering major at UCLA, and I wanted to express my interest in this project!

I've worked with agentic systems in the past through the Gemini CLI, where I implemented task-tree visualizations for the coding agent's tool calls/subagent calls (though it was unfortunately never merged). I believe a lot of the experience gained from this work translates directly into this project. I've also taken ECE147 and ECE239AS, UCLA's core undergrad/grad deep learning classes, and joined the professor's lab as a result (working on BCI robotics applications). Additionally, I've owned full-stack projects end-to-end as a software developer lead in Nova, UCLA's premier nonprofit software development club, so I'd be more than happy to create the UI for the Agentic AI component :)

It seems like all the Good First Issues of this repo is claimed, so I was wondering if you had any prerequisite tasks for me to complete, as well as any advice for my project proposal? I also have a couple of questions regarding the project:

1. Which components of the Edge Manageability Framework would the agent primarily interact with — Edge Infrastructure Manager, Edge Application Orchestrator, or across multiple? Would it call their existing REST APIs, or would new interfaces need to be built?
2. Would the LLM component run locally via OpenVINO (e.g. an on-device model), or is it expected to call an external inference endpoint? Is keeping everything on-device a hard requirement?
3. Is the dashboard expected to be integrated directly into the existing orch-ui, or built as a standalone agentic AI UI alongside it?

Lastly, what would be the best first issue or component to familiarize myself with, would you recommend starting with the EMF repo or a specific orchestrator service?

Thanks for your time!