What is actually being scored
By the end: You know what the interviewer is looking for in a round where you are expected to use an AI tool.
- Why this round exists
- What interviewers say they score
- The fastest way to fail it
- The seniority inversion
New to working with coding agents? Writing software alongside a coding agent — a model that reads your repository, proposes changes across several files and runs commands — rather than typing every line. The skill is not the wording of the request. It is cutting the task small enough that the result can be checked, reading the diff line by line, and recognising early when it has gone somewhere wrong. Why it matters
A newer kind of coding round, where you are handed a large task and expected to use an AI agent rather than avoid one. What confuses candidates is that finishing it is not the point, because a good agent will finish it for almost anyone. This section is about what is on the rubric instead, how you scope, plan, review and verify, and about the quickest way to fail, which is accepting a diff you have not read.
A newer round, with different rules#
Some companies now run a coding round where using an AI coding agent is expected rather than forbidden. The task is usually larger than a traditional exercise, because the assumption is that you will move faster.
The important thing to understand is that the output is not what is being scored. Anyone can produce working code with a good agent. What is being scored is your process.
What interviewers report scoring#
- Scoping. Did you work out what the problem actually is before generating anything?
- Planning. Did you iterate on an approach before executing, or accept the first suggestion?
- Delegation. Did you break the work up sensibly, rather than asking for everything at once?
- Review. Did you actually read the diff, or accept it because it looked plausible?
- Verification. Did you write or run tests without being told to?
The fastest reported way to fail
Paste the prompt, wait for the agent, glance at the output, say it looks good. Interviewers describe this as the clearest negative signal in the round, because it demonstrates no judgement at any point.
The ending nobody prepares for#
Many interviewers finish by taking the tool away and asking you to walk through the code: why this structure, what you rejected, where it would break, what you would change. You are expected to answer confidently without the agent.
This is the part that separates candidates. If you did not read what was generated, you cannot do it, and the whole round unravels at the end regardless of whether the code worked.
The seniority inversion#
Hiring managers running these rounds consistently report that more senior candidates use the tool less, not more. Junior candidates tend to one-shot the problem and then cannot explain any of it. Senior candidates delegate narrower pieces, review harder, and keep hold of the design.
So leaning harder on the tool to look modern works against you. Use it where it is genuinely faster and own everything it produces.
Worth knowing about this round
Some engineers consider it poorly defined, and interviewers running it do not always agree on the bar. That is a fair criticism and it does not help you in the room. Prepare for the rubric as reported and be ready for the expectations to be vague.
Where people go wrong
Treating the round as a speed test. Finishing early with code you cannot defend scores worse than finishing less with clear reasoning throughout.
If this comes up in an interview
How do you use AI tools when writing code?
For scoped, well-understood pieces where I can review the result quickly: boilerplate, test scaffolding, a function with a clear contract. I keep the design and the structure myself, read every diff, and run tests before accepting anything. I do not delegate work I could not have written.
How do you make sure AI-generated code is correct?
Scope the request narrowly enough that the output is reviewable, read the diff rather than the summary, check it against the actual requirement rather than plausibility, and run tests. For anything non-trivial I write the test first so there is something objective to check against.
When do you not use an AI agent?
When I do not yet understand the problem, because I would be delegating the thinking rather than the typing. Also on anything security-sensitive or subtly stateful, where a plausible-looking wrong answer is expensive and hard to spot in review.
Build this: Do a timed exercise with an agent, narrating your reasoning out loud the whole way. Then close the tool and explain every decision from memory. The gap between what you can explain and what you shipped is what the round measures.
Watch it explained
This is one section of Working with coding agents, 7 hours across 2 sections. Read it in full to get the whole topic in order.
Stop 39 of 50 on the AI Engineer route
0 of 50 topics done · 0h behind you, 603h to go
What you have marked done is stored in this browser only. It does not reach a server, and clearing site data will clear it.