Order matters
A reviewer gives a project roughly ninety seconds before deciding whether to open the repository. Problem, constraint, outcome, then stack — in that order.
Depth over count
Five tutorials rebuilt from a video are indistinguishable from each other. One system with real constraints, a decision log and a description of what broke is a conversation for an entire interview.
- State the constraint that made the problem non-trivial.
- Show one measurement: latency, cost, error rate, throughput.
- Name a trade-off you accepted and why it was acceptable.
Admit the wrong turn
A documented mistake and its correction is the clearest evidence of engineering judgement anyone can put in front of a reviewer.