Scored criterion by criterion
A working engineer scores each applicable criterion and writes up what held and what did not.
How scoring works
Every score comes from the same five criteria, weighted the same way every time.
The five criteria
Each is scored 0 to 100. Your round score is the weighted average.
- 35%
Technical depth
Whether you know it past the surface answer.
- 25%
Problem solving
How you move when you do not know the answer.
- 20%
Communication
Whether someone else can follow your reasoning.
- 12%
Behavioural
Ownership and trade-offs in situations you were in.
- 8%
Code quality
Naming, structure and correctness of code you write.
Two kinds of round
Scored as a whole transcript
The whole transcript is read once and given a single score out of 100. If scoring cannot run the round is left unscored, never guessed — you keep the transcript and it counts towards nothing.
Not every criterion applies to every round
A criterion that does not apply is removed, not scored zero — its weight is shared across the rest, so every round is still marked out of 100.
| Round type | Scored on |
|---|---|
| Technical / coding | Technical depthProblem solvingCommunicationBehaviouralCode quality |
| System design | Technical depthProblem solvingCommunicationBehavioural |
| Behavioural | Problem solvingCommunicationBehavioural |
| HR / other | CommunicationBehavioural |
How the leaderboard combines them
70%live interview average30%AI practice average
- One kind of score is used on its own. No AI rounds does not drag your average down — the blend applies only when you have both.
- One rated session to appear. A single session is a one-point sample, so early positions move sharply.
- Ranked on that score alone — not curved against anyone else.
The leaderboard is opt-in and you can leave any time from your profile. Individual scores and transcripts are never public.