System Design · Microsoft · Medium
Design a platform that lets operators oversee a fleet of lab machines executing compute jobs. Each machine reaches out on a regular schedule to ask whether there is something for it to do. An assignment names a compute payload that the machine has to fetch, check for integrity, launch, and then keep an eye on while it runs. A central dashboard should tell operators which stage each payload is in: downloading, running, completed, crashed, or seemingly hung. When a payload…