Marauder Bot

The invitation email lands: "Complete your assessment within 7 days." Which platform it links to changes what you should expect and how you should warm up. Here's a field guide to the formats you're most likely to meet, and the preparation that transfers across all of them.

HackerRank

The highest-volume player, especially in big-company and enterprise pipelines. A typical screen: one to three coding problems plus a multiple-choice section, 60–90 minutes, in-browser editor with visible sample tests and hidden final tests, partial credit from passing a subset of test cases.

Know before you go: hidden tests lean on edge cases and maximum-size inputs, so read the constraints line and mind your complexity. Partial credit makes "correct but naive, submitted" worth real points. The site's practice section uses the same editor as real assessments — warm up there so the environment costs you nothing on the day.

CodeSignal

Known for standardized, score-based assessments (results are sometimes shared across companies, so one strong sitting can serve several applications). Typically four problems in ~70 minutes, ramping from warm-up to genuinely hard, with scoring that factors in test-case passage across attempts.

Know before you go: the ramp rewards triage — bank the early problems fast to buy time for the back half. Because scores can be reused, treat a CodeSignal sitting as higher-stakes than a single-company screen; don't take it cold.

Codility

Common in European pipelines. Fewer problems, more algorithmic depth, and an explicit emphasis on correctness and performance: solutions are scored against both functional tests and large-input performance tests.

Know before you go: a solution that's correct but quadratic will visibly lose points here. State-of-the-art pattern knowledge (our practice workflow builds it) pays off more on Codility than anywhere else.

Company-built assessments

Plenty of companies run their own: a timed form, a shared editor, a take-home in a repo. Formats vary too much to generalize, with one constant — instructions matter more, because conventions don't carry over. Read the rules twice: time limits, allowed resources, submission mechanics.

Proctored variants

Any of the above can arrive wrapped in proctoring: webcam recording, screen recording, browser lockdown, ID checks. The invitation email discloses it. Two notes: set up your environment early (proctoring tech eats setup time and occasionally refuses to work — our time-management playbook covers the pre-flight checklist), and know what you're agreeing to before you click accept. Where you've agreed to work unassisted, honor it — our position on that is public.

What transfers across every platform

Formats differ; the fundamentals of doing well don't:

The material is the same. Every platform draws from the same well: algorithm problems and CS-fundamentals questions — the three question types in different wrappers.
Pacing is the meta-skill. Triage pass, per-question budgets, bank the sure points, never leave an MCQ blank.
Practice in the real format. At least one full timed session in each platform's own environment before the real sitting. Editor quirks, test-visibility rules, and submission mechanics are free points for the prepared.
Review every practice session while it's fresh. This is where most candidates leak value — a finished practice assessment is a goldmine only if you actually mine it. Marauder Bot makes that step nearly free: open the practice page, run one analysis, and get structured answers for everything on it — working code with unit tests for the coding problems, answers with explanations for the MCQs — even on platforms that disable copy-paste or render questions as images. Attempt, analyze, compare, move on.

Whichever platform your invitation names, the loop is the same: learn its format, practice in it once or twice, review hard, and walk in knowing the environment. The free trial (15 page analyses over 7 days) is sized for exactly that week of preparation.