Session abstract
What you’ll learn
What happens when a user starts a report, import, or other multi-step job and leaves before it finishes? In many web apps, the work is tied to the HTTP request that started it. The browser closes, the request ends, and the user has no status to check or result to retrieve. In this 90-minute hands-on workshop, you will deploy an app built that way and close the tab during a run. Then you will change the backend and will break the work into sequential and parallel tasks, inspect the run, follow its status, and recover the result after reopening the page. You will leave knowing when work belongs inside a request, when it needs a job of its own, and how long-running jobs fit behind modern web applications, including AI agents that coordinate multi-step work across tools, retries, and parallel tasks.
