Real code, real tests
Write Python or JavaScript in a full editor and run it in a sandbox against hidden tests, including large inputs that catch slow solutions.
Mock interviews, on demand
Practice the interview, not just the problems.
Talk through your approach with an AI interviewer that asks follow-ups, run your code against hidden tests, and get a scored debrief. Then run a full loop and find out what level you interview at.
- Coding problems
- 150
- Topics
- 15
- Behavioral questions
- 64
- Levels
- Junior to Staff
Everything a real interview asks of you
AI-assisted interviews
Some companies now allow AI in coding rounds. Practice with an assistant that, like real ones, sometimes slips a bug into its code, and get graded on catching it.
System design
Write a design doc that turns into live diagrams while a Socratic interviewer walks you from requirements to scaling and trade-offs.
Behavioral, with a story bank
Answer questions that probe for your own actions and measurable results. Every story you tell is saved so you can refine and reuse it.
Full interview loops
Run a timed sequence of rounds at a target level and get a hiring-committee verdict with the evidence from each round.
An interviewer that pushes back
Pick a supportive, balanced, or strict interviewer, by text or voice. It asks follow-ups, notices hand-waving, and debriefs you honestly.
Interview loops
Find out what level you interview at
Pick Junior, Mid, Senior, or Staff. Take the rounds a real loop has (coding, system design, behavioral), each on a 45-minute timer. A hiring committee then weighs every round against that level's bar and tells you hire or no hire, the level the evidence supports, and what to practice next.
Start a loop →
Behavioral
Stories that hold up to follow-ups
The interviewer keeps asking until it hears what you did and what changed because of it. Feedback scores every answer on situation, task, action, and result, and rewrites your weakest one from your own facts.
Problem library
150 original problems, verified by running them
Fifteen topics from arrays and hashing to graphs and dynamic programming, at Easy, Medium, and Hard. Every problem's tests are checked against two reference solutions and a brute force, so a failing test means your code, not the problem.
Questions
Do I need an account?
No. Open the app and start. Sign in with email, Google, or GitHub whenever you want to keep your history across devices; everything you did before comes with you.
Which languages can I use?
Python and JavaScript, for both writing and running code.
Are these real interview questions from companies?
No. The 150 coding problems are original, and every expected answer is computed by running a reference solution, then checked against a second solution and a brute force.
Why would the AI assistant give me buggy code?
Because real assistants do, and AI-allowed interviews test whether you check their output. The app tells you up front; the debrief shows which bugs you caught.
Is my code or interview private?
Yes. Sessions stay in your browser, or in your account if you sign in. Nothing is public unless you create a share link, which shows only scores or a verdict, never your code or transcript.
How do I know I'm improving?
The progress page rates you per topic and per skill, weighting recent interviews more, and charts your scores over time. Problems that went badly come back for review after 1, 3, 7, and 21 days until you pass them well.
How accurate is the verdict?
It is AI feedback calibrated against written bars for each level, not an employer's decision. Use it to find your weak spots before the real loop.
Your next interview, before the real one
Start with one problem. No signup, nothing to install.
Start practicing free