Timed DSA rounds shaped like real company OAs. Build your round, solve under the clock, and get coached on every decision, including the ones between problems.
4
Round shapes
4
Languages
100%
Verified problems
def longest_window(events, limit):
left, best, seen = 0, 0, {}
for right, e in enumerate(events):
seen[e] = seen.get(e, 0) + 1
...
Every shape you can build here mirrors a real online assessment. Pick the one your target companies actually run.
2 problems
Easy
Medium
The fast screen. Bank the easy one, then earn the medium.
2 problems
Medium
Medium
Two real problems, no warmup. Pacing is the whole game.
4 problems
Easy
Medium
Medium
Hard
The full ramp on a tight clock. Every minute is a decision.
4 problems
Easy
Medium
Medium
Hard
The full ramp with room to think. Depth still gets scored.
Pick a language, pick how many problems, pick the mix or the time. Every shape you can build mirrors a real online assessment.
A full-screen editor, a countdown, and a problem rail. Run against the visible tests, then submit to lock each problem in. Hidden tests only fire at submission.
An AI judge reads your code, your timing, and the order you tackled questions in, then writes per-problem coaching and a time note for the whole round.
Every round ends with an overall score, four sub-scores, per-problem coaching, and a time note. Here is how the score is built.
Difficulty-weighted hidden test results, adjusted for the quality of your approach.
Judged against your chosen shape's budget and the ramp, never an absolute clock.
Boundaries, empty inputs, off-by-ones. The failures that pass the happy path.
Naming, structure, dead code. Tests are the floor: green code still gets quality coaching.
The round records when you opened each question, how long you stayed, and when you bailed and came back. The judge tells you where the minutes went and what the winning move was.
Visible tests are for iterating. Hidden tests run only when you submit, and results stay sealed until the round ends.
Every draw excludes the problems from your last three rounds, so a retake is always a real retake.
Scores never mix across shapes. A 45-minute sprint and a 90-minute ramp are different disciplines, and your progress charts treat them that way.
Every problem ships with a reference solution that is machine-verified in your language before it can ever be served to you.
When time runs out, whatever you have is collected and scored, exactly like a real OA. No pause button, no overtime.
Free to start. The clock is waiting.