SEMINARS

The Intelligent Codesign Cloud Lab for Non-Equilibrium and Low-Dimensional Materials Programmable Cloud Lab (ICoN-PCL) seeks to accelerate material-product co-design for next-generation electronics and metallic alloys for extreme conditions. A core premise of ICoN-PCL is that autonomous research is only as good as the workflows and data underneath it. Every experiment and simulation should be findable, accessible, interoperable, and reusable (FAIR) by default, not as an afterthought. That is how the result of one experiment can inform the next decision, whether a human or an AI agent makes it, and whether it happens in our lab or at another node.

ICoN’s digital partner, nanoHUB, has served more than 200,000 simulation users across 800+ US institutions. We will describe the nanoHUB components that make FAIR workflows practical:

• Sim2Ls: simulation tools with declared, typed, and verified inputs and outputs that can be called from code. Their results and provenance are indexed automatically in a results database you can query.
• Cloud execution with federated identity: researchers at any institution can publish new Sim2Ls and run workflows without installing anything locally.

We will then show how these resources become available to AI. Our MCP servers give LLM agents access to nanoHUB tools, the results database, and external resources. AURA, our AI research assistant, reasons over tool and data metadata to help users ideate, plan, and carry out research tasks. It can query existing results before spending new compute, launch simulations, and record each step for others to inspect.

The session will include short demos and we invite all PCL nodes to connect their agents to our MCPs, publish their own workflows as Sim2Ls, and work with us on shared ways to describe workflows and results across nodes.