Design a Job Scheduler
This came up in an Airbnb software engineer onsite in April 2026, with the candidate quoting it as a job scheduler "at roughly 10M scale, used internally." They called it "a staff-level design problem," so prepare for a senior-and-up deep dive whatever level you're interviewing at. An August 2026 onsite narrowed the same idea to one-time payment jobs. That version leans on durable claiming, retries, races between cancel and execute, reconciliation with the payment provider, and idempotency across client, worker and provider failures. A scalable job scheduler also appears on older Airbnb question lists.
Picture the jobs that live on a platform like this. Release a host's payout a day after check-in. Expire an unpaid reservation hold after 24 hours. Send a review reminder the morning after checkout. Rebuild a search index every night at 2 a.m. Most are tiny. Some move money. Each one has an owner who wants to know it fired on time, and only once. Running one job at the right time is easy. The interview is about the failures. A scheduler can crash right after it decides to run something. A worker can die halfway through. The queue can deliver the same message twice, and the network can cut a worker off from the database while it's still working. The design has to survive all four without running anything twice.
AirbnbPremium lesson · System design round
The Airbnb design, worked end to end.
- Requirements and scopeSystem design · Frequently asked
- The API and data model
- The design, drawn out1 diagram
- 6 deep divesWith the tradeoffs
$49 a year · all 107 companies · instant access
20,000+ candidate reports · guide updated
Airbnb interview guideKnow what Airbnb asks before you walk in.
24 coding · 6 system design · 20 behavioral questions
Get their questions first.
InterviewLegend is independent and not affiliated with or endorsed by any company it covers.