Interviewed for a Staff SWE role at McDonald's Global Technology in early 2026. Didn't take the offer (comp gap, longer story) but the system design round was actually one of the better ones I've done in this search cycle, so worth documenting.
Setup: 60 minutes, two interviewers (one tech lead, one senior engineer). Google Meet with a shared digital whiteboard.
The prompt they gave me: Design the McDonald's mobile ordering backend. Handle concurrent orders, handle restaurant capacity, handle failure gracefully if a restaurant's local connection drops.
On the surface it sounds gimmicky. In practice it's a genuinely good system design question because it has real constraints: High read volume (menu browsing) vs. moderate write volume (order placement) Hard latency requirements: user shouldn't wait more than 2 seconds to confirm order Partition tolerance matters: a restaurant losing internet should not brick the whole ordering system They have actual proprietary POS hardware in-store, so your design has to account for an edge compute layer
What they probed on: I started with a standard API gateway + order service + restaurant service topology. They pushed me on: what happens when the order reaches the restaurant and the POS is offline? How do you reconcile? Do you use an event-driven approach or is polling acceptable here?
I went with event-driven (Kafka, or similar), offline-capable edge device with local queue that syncs on reconnect. They seemed to like that I named the failure modes before they asked.
What they didn't care about: My specific choice of database. They asked why I picked Postgres over Cassandra but weren't looking for a single right answer. The reasoning mattered more.
For staff-level candidates: They wanted to see me drive the conversation, not wait to be led. They also asked one 'how would you lead the team implementing this' question at the end, which felt like a bar check on the L-adjustment.
Prepare one solid 'distributed system with edge compute' design. That's the flavor of problem they have.