We were migrating a core service to a new data store and halfway through, two engineers left the team and our timeline got cut by six weeks. I picked up their work, reorganized the remaining tasks, and kept the project moving. I worked closely with the PM to reprioritize scope, we cut a few lower-priority features, and we shipped on the revised deadline. The service is running in production and the team was really relieved we made it. It was a tough situation but we pulled through.
We were migrating a high-traffic order routing service to DynamoDB. Six weeks before launch, two engineers left and leadership asked us to absorb their scope without slipping the date — this service processed 40 million requests a day and a delay meant downstream teams couldn't ship their Q4 features either. I triaged every remaining work item, owned the two riskiest modules myself, and built a lightweight load-testing harness that let us validate correctness in two hours instead of two days. We shipped on schedule. Post-launch p99 latency dropped 34 percent and zero downstream teams were blocked. I documented the harness as a reusable runbook — three other teams have since adopted it.