Auto-generated portfolio that updates itself as a person keeps working

A rendering system, part of a training platform, that builds a portfolio for each person from their real work on that platform: submissions, conversations, grading verdicts, and their overall progress story. When someone is preparing for two different kinds of roles, the system can render two different, coherent portfolios from that same underlying work instead of one blended, averaged-out summary.

When someone is preparing for two different roles, they need two different portfolios that tell two different stories from the same underlying work. And a portfolio like this needs to be alive, updating itself as the person keeps working, not a one-time snapshot that goes stale the moment they finish their next task. A static portfolio is fine for a snapshot in time. Someone who is still actively working needs one that moves with them.

Two database tables (profile and assessment) keyed by person-plus-role-track, so two role tracks render as two coherent rows. A background workflow runs the refresh in four steps: gather the person's inputs, write the profile narrative, write the assessment, and save both. The trigger lives upstream: every time a piece of work finishes being graded, a background job kicks off this refresh automatically, carrying what just happened as context. So the portfolio updates within seconds of the person finishing a conversation, a submission, or a session, without the grading step ever having to wait on it. A lazy-seeding step picks a starting template based on the person's stated goal if no portfolio exists yet, so their very first session already has some shape behind it.

Co-contributor on the per-completion refresh trigger, the input-gathering step, and the wiring that connects the skill-tracking system to the portfolio refresh.