How do you turn a recruiter's description of each upcoming round into a preparation plan?

## From a stage list to a plan Most candidates hear the stage list, feel reassured that a plan exists, and then prepare the same way they would have anyway. The value is in the conversion: the recruiter has just told you what each round is measuring, and that should visibly change what you do with the weeks before it. The conversion is a table. One row per remaining round, four columns: | Round | Audience | What it assesses (their words) | What I will practise | |---|---|---|---| | Practical exercise | Two engineers | Reasoning out loud on an unfamiliar problem | Working through problems while narrating, on a timer | | Threat-modelling round | Security engineers | Structured decomposition of an unfamiliar system | Modelling systems I do not know, from a public description | | Incident walkthrough | Senior engineer plus manager | Depth and honesty on something I actually ran | Two incidents rehearsed with real timelines and numbers | | Team round | Hiring manager | Fit for the scope of the posted level | Examples pitched at that scope, not at a smaller one | Writing the third column **in the recruiter's words** matters. "Reasoning out loud" and "a correct finished artefact" are different assessments that produce different rehearsal, and the difference is only visible if you kept the phrasing instead of collapsing it to "technical round". ## Weighting the rows Rounds are not equal, and evenly spread preparation is the default mistake. Two signals tell you where the weight sits. First, **count**: if two of five stages are domain rounds, the loop is telling you plainly that the domain is what it cares about. Second, **audience seniority and duplication**: a subject probed twice by different people is a subject the decision turns on. So in the example above -- five stages across roughly seven weeks, two of them domain rounds on the same subject area -- most of the rehearsal budget belongs to threat modelling and incident depth, and comparatively little to general practice. That is a decision you can only make because you asked what each round assesses. ## Sorting rounds by mode Rounds fall into two rehearsal modes, and confusing them is why preparation sometimes feels wasted. **Live-performance rounds** assess how you think in the moment. They are prepared by doing the thing out loud, under time, ideally with someone listening. Reading about it does almost nothing. **Retrieval rounds** assess work you have already done. They are prepared by getting your material straight: which examples you will use, what actually happened, the real timeline, the real numbers. The failure here is not lack of skill but a hazy memory of your own work -- being unable to say how long the remediation actually took, or how many services were affected, when those are precisely the details that make an account credible. Most loops mix both, and the mix is exactly what the recruiter just described to you. ## Converting vague labels Sometimes you get a label rather than an assessment -- "a culture round", "a technical screen". Two questions convert it, and both are fair to ask a recruiter: - *What does a strong candidate do differently in that round?* Recruiters hear debrief language repeatedly and can often reproduce it, even without doing the work themselves. - *Is it a discussion of past work, or something worked through live?* This is the mode question, and it is usually answerable straight away. If the recruiter genuinely does not know, that is worth recording too: a round nobody can characterise is one to prepare defensively, with your examples in order. ## The last column Each row should end with the question you plan to ask that audience -- the ones you deliberately did not spend on the screen. Arriving at the domain round with a question written for security engineers, and at the manager round with a different one, is the visible output of having planned rather than improvised. ## Where this goes wrong Three failures are common. **Flat preparation**, ignoring the weighting the recruiter handed you. **Preparing for the label rather than the assessment** -- practising puzzles for a round that is actually about walking through past work. And **letting the plan rot**: the timeline is a real constraint, and a plan built for roughly seven weeks needs re-cutting if scheduling compresses the last three rounds into one week. Ask the recruiter to flag that if it happens; it is exactly the kind of thing they can tell you.