Why this site exists
A cloud interview rarely asks what a service does. It asks what happens when the service fails, who is allowed to call it, what it costs at ten times the load, and what you would do differently next time. Those are judgment questions, and lists of service definitions don't prepare you for them.
The work itself has also changed. Cloud teams now run language models and agents as production workloads. The same engineer who designs a VPC is asked how an agent should authenticate to a ticketing system, where its generated code runs, and who notices when one request turns into forty model calls. Interview loops are starting to reflect that, and preparation material has not caught up.
CloudInterviewPrep is our attempt to fix both problems in one place. The goal is ambitious: to become the best cloud interview preparation resource in the world. That is a standard to measure ourselves against, not a claim we expect you to take on trust.
Who it is for
- Cloud engineers preparing for hands-on infrastructure and troubleshooting rounds.
- Solutions architects preparing for whiteboard design and migration conversations.
- Platform and DevOps engineers preparing for Kubernetes, CI/CD and multi-tenant platform questions.
- Site reliability engineers preparing for SLO, incident and capacity discussions.
- Cloud security engineers preparing for identity, encryption and threat-modelling rounds.
- Customer engineers, sales engineers and forward deployed engineers who design with a customer in the room.
- Engineers moving into AI infrastructure who need to reason about agents, model endpoints and GPU capacity as cloud workloads.
What it covers
The material is organised into eight tracks that mirror how a cloud interview moves between layers:
- Cloud architecture and system design
- Compute, containers and Kubernetes
- Networking and traffic
- Identity, security and compliance
- Data and storage
- Reliability and observability
- Cost and FinOps
- Agentic systems and AI infrastructure
Start with the 40 worked cloud interview questions, then go deeper with the cloud system design framework and the agentic systems interview guide.
Editorial standard
Every answer on this site is written to be checked, not just read. In practice that means:
- Assumptions are stated. An answer says what load, budget and failure model it assumes, so you can see when a different assumption would change it.
- Concepts come before products. We explain the idea first and name AWS, Google Cloud or Azure services second, because the idea is what transfers between interviews.
- Volatile facts carry a date. Service names, limits and prices change. Where we state one, we say when it was true.
- No invented interviews. We never claim a question was asked at a named company, and we never publish fake reviews, success rates or view counts.
- Tradeoffs over verdicts. Where reasonable engineers disagree, we show both sides and the condition that decides between them.
Cloud moves quickly and we will sometimes be wrong. Corrections are the most valuable feedback we receive, and fixing them comes before publishing anything new.
Independence
CloudInterviewPrep is independent. It is not affiliated with, sponsored by or endorsed by Amazon Web Services, Google Cloud, Microsoft or any employer we mention. Product and company names are trademarks of their respective owners and appear only to explain concepts.
What is next
The site is growing track by track. Coming additions include deeper worked cases for each track, practice questions with follow-ups that change a constraint, and interactive exercises for estimation and failure analysis. Everything published today is free to read, with no account and no email required.