Strong interviews test more than recall.They test what survives"why?"
An AI examiner that challenges each step of your reasoning and follows up when an assumption is weak. Practice with timed mocks and live, rubric-graded problems.
$30 monthly · $144 every 6 months · $399 lifetime
Define conditional probability.
DEF · Probability · ~3 min · 2 hints available
Define P(A | B). State the condition under which it is defined.
P(A | B) = P(A ∩ B) / P(B).
You gave the formula. What condition must B satisfy, and what does conditioning on B mean?
The problem bank includes questions attributed to interviews at
A familiar solution can still fail under a small change.What matters is whether your reasoning adapts.
Pattern practice builds speed.It does not test every assumption.
Under pressure, one changed condition can expose a gap.A mock makes that gap visible.
Define. Understand. Apply.
A layer is marked proven only when its rubric criteria are defended — not when you simply finish the drill.
- DEF
Can you actually define it?
State it precisely. Then defend every word.
ExaminerDefine a limit of a sequence.
YouA sequence aₙ converges to L if for every ε > 0 there exists N such that |aₙ − L| < ε for all n ≥ N.
ExaminerWhat does "for every ε > 0" actually mean? Why not just "small ε"?
- UND
Can you explain why it works?
Explain why it works. Derive, don't recall.
ExaminerWhy does the chain rule work? Don't compute—explain.
YouBecause the derivative is a local linear approximation. If f locally scales by f′(g(x)) and g locally scales by g′(x), composing them scales by the product. The chain rule is just "linear approximations compose by multiplication."
ExaminerRight—now where does that argument actually fail?
- PRB
Can you reason through it under pressure?
Solve under pressure. Every step justified.
ExaminerA drunk man takes n steps on a line, each ±1 with equal probability. What's his expected squared distance from the origin?
YouLet Sₙ = X₁ + ⋯ + Xₙ where each Xᵢ is ±1 with prob ½. Then E[Xᵢ] = 0 and E[Xᵢ²] = 1. So E[Sₙ²] = E[(ΣXᵢ)²] = Σᵢ E[Xᵢ²] + Σᵢ≠ⱼ E[XᵢXⱼ] = n + 0 = n.
ExaminerYou dropped the cross terms. Justify it.
Run the mock. Review the gaps. Drill them.
Start with a timed mock. Its debrief records the questions and criteria that held — and those that did not.
- 01
Configure a timed mock
Set the interview target and choose two to four questions. The timer adjusts to seniority and question count.
- 02
Work without hints or pauses
Questions unlock one at a time, and each submitted answer stays on the record.
- 03
Review the recorded gaps
See the per-question verdict, missed criteria, and full session debriefs.
- 04
Drill, then run another mock
Use the problem bank or curriculum to work on those gaps, then repeat the sitting.
Timed. No hints. No restart.
Choose two to four questions. The time limit adjusts to seniority and question count, and the examiner stays in interview mode throughout.
The clock starts immediately. No pause, no restart.
A completed sitting ends in one aggregate debrief.
Questions unlock one at a time.
A verdict backed by recorded criteria.
Completed, scored sessions record a verdict, the rubric criteria you defended, and the criteria that remain unresolved.
One question broke; two held.
Question 1 was not defended
The debrief opens the first question that was not defended and links to its full session record.
319 problems. Drill any of them live.
Browse by pillar, layer, or interview style. Every listed problem has a reviewed rubric and can run live against the AI examiner.
Best-of-Seven Series
Google · Probability
Runs live · rubric-graded
Unfair Coin Detection
Probability fundamentals
Runs live · rubric-graded
Russian Roulette Strategy
Brainteasers
Runs live · rubric-graded
What practice alone may not reveal.
Usually shows you a solution, then leaves the next step to you.
Records the missed criterion and keeps probing the reasoning behind it.
You infer readiness from completion and confidence.
A layer is marked proven only when its rubric criteria are defended.
Often the final answer or whether you completed the question.
The reasoning behind each step, against a problem-specific rubric.
A memorized pattern may not transfer to the changed assumption.
Follow-up probes ask you to explain how your reasoning must adapt.
A list of completed questions and your own notes.
A saved attempt, verdict, missed criteria, transcript, and debrief.
Generated dialogue. Locked grading.
Published problems and rubrics are hand-reviewed. The AI phrases the dialogue, while deterministic rules control scoring and session state. Every generated reply is checked before display.
- 1
Hand-reviewed.
Every published problem and rubric is reviewed before it appears in the bank.
- 2
Rubric-locked.
The scoring criteria stay fixed during a session. The model may vary its wording, but it cannot rewrite the grading rules.
- 3
State-controlled.
The model proposes an assessment; deterministic rules decide rubric updates, the next move, and session completion.
- 4
Checked before display.
Each generated reply passes an output guard. A rejected reply is replaced with a safe probe.
- OutputReaches you
“Why log? Why specifically that function?”
Round three at WorldQuant. I’d spent four months preparing for it.
When he asked me to explain logistic regression, I relaxed a little. Finally, something familiar.
I started answering.
Then he cut in:
“Why the log?”
And my brain just… froze.
Not because I hadn’t used logistic regression before — I had, a lot. But I’d only learned how to explain it well enough to pass interviews, not deeply enough to defend it.
I remember sitting there in silence, trying to think of anything to say.
Nothing came.
I knew right then the interview was over.
That moment changed how I think about interview prep. Most people practice polished answers. Great interviewers look for the cracks underneath them.
That’s why I built the thing I wish I’d had back then.
Choose the access that fits your timeline.
Subscribe monthly or every six months, or pay once for lifetime access. Every paid option includes the current four-pillar curriculum, live sessions, timed mocks, and mastery tracking.
Lifetime
One payment. Full access for the life of the product.
- Current four-pillar curriculum
- Unlimited live sessions + timed mocks
- Mastery tracking across sessions
- No subscription, no renewals
Founding
Full access while your subscription is active.
- Current four-pillar curriculum
- Unlimited live sessions + timed mocks
- Mastery tracking across sessions
- Direct line to the founder
Enterprise
For teams, bootcamps, and universities preparing candidates at scale.
- Custom quote based on cohort size
- Discuss onboarding and problem-set needs
- Direct contact with the founder
Asked before buying.
Candidates preparing for rigorous technical interviews — quant trading, ML, research, math grad school. Self-learners who can recognize a problem but stall when it’s rotated 30°.
The real interview won't go easy on you.Neither will this.
Start a timed mock and find the gaps before the real interview.
Subscriptions can be canceled from billing settings.