Interview scorecard
Backend Engineer
Score each skill from 1 to 5. A 1 is below the bar for this role, a 3 meets it, and a 5 is above it. The examples show what each sounds like. Use 2 and 4 for answers that fall in between.
API design
Designs clear, consistent interfaces that other developers can use safely and that can change without breaking them.
Questions
- Tell me about an API you designed. Who used it and what did they need?
- Tell me about a time you changed an API that others already used. How?
- How do you decide what errors an API returns, and how do you describe them?
1: Below the bar
Describes endpoints without thinking about users. Changes break callers, and errors are vague or inconsistent.
3: Meets the bar
Designs consistent endpoints with sensible names and errors. Knows changes can break callers and avoids it with some care.
5: Above the bar
Starts from what callers need, keeps naming and errors consistent, and plans versioning and a safe rollout before changing anything.
Data modeling
Shapes data into tables and relationships that fit how the application reads, writes and changes it.
Questions
- Tell me about a data model you designed. How did you decide on its structure?
- Tell me about a time a schema had to change after launch. What happened?
- How do you decide what to store together and what to split apart?
1: Below the bar
Lists tables without explaining choices. Doesn't consider how the data will be queried or how it might change later.
3: Meets the bar
Explains the main choices and how data is queried. Handles a schema change adequately but gives limited thought to future growth.
5: Above the bar
Ties every choice to real query patterns, names trade-offs such as duplication versus joins, and plans safe schema changes in advance.
System design
Designs services that stay correct and available as load, data and failures grow.
Questions
- Tell me about a service you designed that other systems relied on. What had to be right?
- Tell me about a time part of a system failed. How did the rest behave?
- Describe a time you traded consistency for speed, or the reverse. Why that choice?
1: Below the bar
Lists components without saying how they fail or grow. Can't explain why data lives where it does.
3: Meets the bar
Describes a working design with clear boundaries. Covers common failures but not load growth or partial outages.
5: Above the bar
Explains boundaries, data ownership and failure handling unprompted, and gives a real trade-off between consistency, cost and speed.
Debugging
Finds the cause of a backend fault by testing ideas step by step and confirming the fix.
Questions
- Tell me about a hard bug in production. How did you find the cause?
- Describe a time your first theory was wrong. What did you do next?
- Tell me about a bug you fixed. How did you know it was truly fixed?
1: Below the bar
Guesses and changes code at random. Can't describe how they confirmed the cause or that the fix worked.
3: Meets the bar
Follows logs and errors to the cause with a clear sequence of steps. Checks the fix, but may not prevent a repeat.
5: Above the bar
Forms and tests specific theories, narrows the cause quickly, and proves the fix. Adds a test or alert to prevent repeats.
Code quality
Writes code others can read, test and change safely, and explains the habits behind it.
Questions
- Tell me about a service you inherited from someone else. What made it easy or hard to change?
- How do you decide what to test in a change, and what to leave out?
- Tell me about a code review where you disagreed with the feedback. What happened?
1: Below the bar
Describes code as working once it runs. Rarely mentions tests, naming or reviews, and treats feedback as a hurdle.
3: Meets the bar
Describes readable code with sensible tests and welcomes review. Gives concrete examples but mostly follows team habits.
5: Above the bar
Shows how they shaped code for later readers, picks tests by risk, and uses review to improve team standards.
Overall recommendation
Decide on your own, before you compare notes with the rest of the panel.