AI tutor · Mock interview
A system design interview, with an interviewer who does not give you the answer.
Pick a real problem and your level. The interviewer asks one question at a time, stage by stage, like a real round. At the end you get a score on ten points, and each score quotes what you actually wrote.
No account: a demo round, your one free question.
Set up your interview
Without an account you can try a demo round: the requirements stage only, about 8 minutes, with a scorecard. It uses your one free tutor question.
The seven stages
- 1. Requirements. Ask what the system must do, for whom, and how fast, big and reliable it must be.
- 2. Estimation. Work out rough numbers: users, requests per second, storage, bandwidth.
- 3. API and data model. Name the main API calls and the tables or records behind them.
- 4. High-level design. Draw the main boxes and how one request travels through them.
- 5. Deep dive and bottlenecks. Find the part that breaks first at scale and fix it.
- 6. Reliability and consistency. Say what happens when a part fails, and what users see after a write.
- 7. Wrap-up. Name the biggest risks, how you would watch the system, and what you would improve next.
The ten points on the scorecard
- Requirements clarification. Asked about features, users and limits before designing.
- Scale estimation. Worked out traffic, storage and bandwidth with numbers.
- API and data model. Named the API calls and how the data is stored.
- High-level architecture. Laid out the main parts and how a request flows.
- Bottleneck identification. Found what breaks first under load and how to fix it.
- Reliability. Planned for failures: retries, replicas, failover.
- Data consistency. Said what users see after a write, and chose a model on purpose.
- Observability and security. Said how to watch the system and keep it safe.
- Trade-off reasoning. Compared options and said why one wins here.
- Communication. Clear, structured answers that are easy to follow.
Questions
How does the AI mock interview work?
You pick a problem from the interview guides and your level. The AI interviewer asks one question at a time, through seven stages: requirements, estimation, API and data model, high-level design, deep dive, reliability and consistency, and wrap-up. It answers your clarifying questions the way an interviewer would, and it does not give you the answers.
How is the scorecard made?
At the end, every one of ten points gets a score from 1 to 5, and each score quotes your own answers as evidence. Before you see it, the site checks that every quote really is in your answers. A score with no real quote is shown as not covered, never guessed. You also get three lessons to read next, all real lessons from the course.
How many mock interviews can I do?
Without an account, one demo round (the requirements stage). A free account gets a short round (requirements and estimation). With either course, a full interview uses 3 of your 10 daily tutor questions when it starts, and every turn inside it is free, so you can do up to three a day.
Can I see my past scorecards?
Yes. When you are signed in, every finished interview is saved on this page with its scorecard and the full transcript.
Want a plan around the mock interviews? Interview in 30 Days puts three of them on days 14, 21 and 28.