How to run a useful mock
Realistic conditions, a fixed rubric, a written debrief, and the same weakness drilled until it moves.
- Simulate the conditions: timer on, camera on, shared editor or whiteboard, no AI assistant (Anthropic's live interviews are "all you", Sources A2).
- Use a partner when you can; otherwise record yourself and score the recording the next day.
- Score with the rubric immediately, then write the one thing to fix next time.
- Repeat weekly and watch the trend in the saved history on each mock page.
- Drill the weakest row between mocks with the matching track lesson.
Every mock page saves attempts in this browser. Scores are 1 to 4 per row: 1 is a clear miss, 2 is below the bar, 3 is at the bar, 4 is a strong signal.
This rubric is ours, built from published signals (for example OpenAI's "well-designed solutions, high-quality code, optimal performance, and good test coverage"). It is not any company's scoresheet. See Sources.
Practice
Written for this site to drill the lesson. Not reported interview questions.
- Book a weekly 60-minute slot for six weeks now.
Check yourself
What should you write after every mock?
The single thing to fix next time.
Saved in this browser.