I want to build an n8n-based AI agent that takes full charge of homework creation, tracking and reporting for an entire class. The goal is to feed the agent both the hard data in each learner’s profile and the softer signals of their preferences, then let it design assignments that truly fit the individual. Here is what I need the workflow to achieve: the moment new curriculum material is specified, the agent should pull each student’s profile from my existing database, weigh it together with their subject interests, learning style and preferred homework format, and generate a set of state-aligned tasks that feel tailor-made. As students complete work, n8n must write progress updates back to the same database and push concise, friendly mobile-app notifications to both the student and the teacher so everyone sees compliance, gaps and growth in real time. I already maintain the database (PostgreSQL) and can expose it through an API. Your part is to stitch everything together in n8n, craft the AI prompts or function nodes that assemble the homework, and make sure the feedback loop runs without manual intervention. If you have prior experience chaining GPT-style models inside n8n, working with educational data standards, or sending notifications through Firebase / Expo, let me know—that will speed things up. Deliverable: a fully configured n8n workflow (exportable JSON) plus clear setup notes so I can drop it into my self-hosted instance and start using it on day one.