Nomad
Six agents. One relocation workflow.
Internship project

Multi-agent systems · Full-stack · AI Micromind
Read the project postThe idea
A relocation workspace built during my AI Micromind training. Six specialized agents bring visa research, documents, housing, budgets, and changing plans into one persistent workspace.
One answer, two destinations
How it comes together
Share the move
Describe the destination, timing, budget and circumstances.
Build a shared plan
Six specialized agents bring research and planning into one persistent relocation workspace.
Follow the moving pieces
Keep documents, housing, budgets, deadlines and progress together.
Adapt the plan
Update the workspace as the destination, timeline, budget or other circumstances change.
More than a single conversation.
Nomad keeps relocation context across conversations so the plan can develop over time. The assistant is the conversational entry point; the journey dashboard makes the resulting priorities, documents and budget visible.
I built Nomad during my AI MicroMind training, working on multi-agent orchestration, persistent data, authentication and the connection between AI reasoning and a full web application.
A closer look at the code
Reading a number from an agent reply
I want an empty budget to stay unknown. This parser rejects empty values before converting numeric strings. From agent-state.ts.
const finite = (value: unknown) => {
if (value === null || value === undefined || value === "" || typeof value === "boolean") return undefined;
const result = typeof value === "number" ? value : Number(value);
return Number.isFinite(result) ? result : undefined;
};Inside the workspace



