It scores every question individually. Hyring Meet produces a technical score with a full per-question breakdown, so you're never left with a single rolled-up number and no way to see which specific answer helped or hurt it. Each question asked during the interview gets its own technical mark, and those individual marks are what any overall score, if you choose to look at one, is actually built from.
To make that concrete: imagine a technical interview with four questions, one on system design, one on a specific algorithm, one debugging exercise, and one about a past project. Under this model, the report would show four separate marks (for example, system design 7 out of 10, algorithm 9 out of 10, debugging 6 out of 10, and project depth 8 out of 10) rather than a single 76 percent figure with no way to trace where that number came from. This example is meant to illustrate the format rather than represent any actual candidate's real scores, since every interview and every set of questions asked is different.
This per-question structure is also what makes per-skill ratings possible in the first place. Each skill rating comes with the number of questions behind it, and that only works because scoring happens at the question level to begin with, rather than being blended into one number from the start.
The full underlying scoring methodology, the exact rubric logic or how partial credit gets handled, hasn't been published in detail, since that's part of how the product works under the hood. If you want to see this in action on an actual sample report with real numbers, the best way is to ask for a demo, which will show you the real per-question breakdown format directly.
