Find out what you'd fail on before the interviewer does.

Practice the questions you'll actually be asked, get graded on what you actually wrote, and see exactly which subtopics would sink you.

Log in Pricing

Free plan, no card. Ten real questions and a real score.

reviewed questions: 10,373
tracks, 12 fully live: 12
subtopics scored separately: 500
tests vs code review: 70/30

Free, no account needed

Learn the basics first, with a real compiler

13 reference courses on the languages, frameworks and databases people actually get hired for, each lesson with an example you run for real and a task you have to solve — graded by the same sandbox that grades an interview answer, with your own mistakes underlined on the line they happened.

Browse the courses

Solution.java

class Point {
  final int x, y;
  Point(int x, int y) { this.x = x this.y = y; }
}

';' expected — line 3

Why not just ask a chatbot?

Four things a chat window cannot do, and a question bank alone will not give you.

Graded on what you actually wrote

Every theory answer is scored against a fixed rubric. Each bullet shows whether you covered it, and quotes the phrase from your own answer that earned the mark. The percentage is arithmetic on that list, never a model's guess at a number.

IQueryable builds an expression tree the provider translates to SQL
From your answer: IQueryable keeps an expression tree that EF turns into SQL
Covered30
Filtering an IQueryable happens in the database
From your answer: the where goes to the database I think
Partial25
Both are deferred until enumeration
Not mentioned
Missing10

Your code actually runs

Tests decide whether it works, in a throwaway sandbox. Seventy per cent of a coding score is tests passing; the review only judges quality.

2 of 3 tests passed2/3

counts_multiple_ordersPassed
ignores_cancelledFailed
```
input: 3 orders, 1 cancelled
actual: 3
```
hidden_1 Hidden testPassed

It remembers what you got wrong

Weak subtopics come back on a schedule until they stop being weak. Strong ones retire. No grinding the same question you already know.

85–100+21d, then +45d, then retired
70–84+14d
50–69+5d
0–49+2d

A score that means something to a hiring manager

Every score is labelled with the level it was measured against, because “68” on its own is meaningless. Questions are hand-written and reviewed, so a wrong premise never wastes your time.

68 / 100 — Needs work for Junior

Know exactly where you would fail

Every answer rolls into a subtopic heatmap. Teal is fine, amber is shaky, coral is the question that ends the interview.

C#74%

SQL Server52%

Indexes & seeks vs scans — 31%. Your code passed 3 of 5 tests but you couldn't explain that a leading wildcard prevents an index seek. Expect a follow-up on this in any SQL Server round.
A risk, worded the way a report words it

C#

Value vs reference types82%
Collections & LINQ74%
async/await & Task61%
Generics70%
Exceptions45%
Delegates & events38%

SQL Server

Joins74%
Window functions52%
Indexes & seeks vs scans31%
Execution plans28%
Transactions & isolation44%
Deadlocks—

0–39
40–69
70–100
Not yet assessed

Questions

Is this just another question list?: No. You answer, and every answer is graded: theory against a fixed rubric that quotes your own words back to you, code by actually running it against hidden tests.
Which stacks are covered?: Languages, frameworks, databases and cross-cutting skills. One plan covers every live track, because full-stack developers should not pay four times.
How long does a session take?: Ten questions is about 25 minutes. Twenty is the usual choice the week before an interview. Thirty is a full loop simulation.
What if the grade is wrong?: Every answer in a session has a 'This grade is wrong' button, and a person reads every dispute.

Your next interview is a week away.

Ten questions, a real score, and a list of what to fix. Free, no card.

Log in Cancel any time from your account. The Sprint simply ends after 21 days.