A company asks you to choose between AWS and Azure for a new workload. How do you make the decision?

The follow-ups they ask next

• The CTO insists on a multi-cloud strategy to avoid lock-in. How do you respond?

  Multi-cloud avoids one kind of lock-in but adds real cost: lowest-common-denominator architecture, duplicated tooling and expertise, and operational complexity that often outweighs the benefit. A senior weighs it honestly, often recommending one primary cloud with deliberate portability at key seams, rather than running everything twice, unless a concrete requirement like regulatory or true vendor risk justifies the overhead.

• How do you keep this decision from locking the company in permanently?

  You cannot avoid lock-in entirely without giving up the managed-service velocity that justifies the cloud, so you manage it: keep portable seams (containers, open standards, abstraction at the few critical boundaries), avoid proprietary services where the lock-in is not worth it, and accept lock-in deliberately where the productivity win is large. The honest senior answer is that lock-in is a trade-off to manage, not eliminate.

What the interviewer is listening for

• Decides from requirements, not brand loyalty
• Weighs existing investments and team skills heavily
• Maps the workload needs to each platform specifically
• Is honest about cost (TCO/egress) and lock-in, and makes a call

What sinks the answer

• Argues by popularity or personal preference
• Ignores existing investments and team skills
• Generic feature checklist with no decisive factor
• Refuses to make a recommendation