CourseAgentic AI Engineer: The Complete Course · Module 15: Runtime and Infrastructure · part 86 of 97
Part 86 · Module 15: Runtime and Infrastructure

Lab: Run the desk as a budgeted background worker

40 min read·13 Oct 2026
Module 15 · Lab · About 80 minutes · Needs: the Module 3 lab's tools.py, lessons 1 (jobs, checkpoints, retries, schedules), 2 (serving, limits, tenants) and 3 (the budget ledger)

What you are building

RETURNS_DESK AFTER MODULE 15 app FastAPI, 202 Redis queue, events run_request start or resume call tools.py cron 08:00, hourly BudgetLedger reserve, settle checkpoints graph_checkpoints.db model
Your new code (blue) takes requests and schedules into a queue, and every run reserves from the ledger (green) before the model (amber) is called; Redis, checkpoints and Module 3's call are infrastructure (grey).

It is Friday 3 July 2026, week 12 (simulated), the last day before full launch. This lab moves the module into returns_desk/worker.py: Lesson 1's RQ jobs, including work_overnight_queue and check_all_waiting for cron_config.py. Every request reserves from Lesson 3's budget ledger before it may call a model. Lesson 2's FastAPI service shares the file, so web and workers share one job definition. Tool calls still go through Module 3's call. You follow Aisha's R-1001.

The rest of this course is yours to keep

This course is bought on its own, once, and stays readable afterwards, including the parts added to it later.