What does the diagnostic give me?

A completed diagnostic ranks weaknesses supported by your recorded answers, separates untested topics from observed risks, and builds a repair plan of targeted drills. It uses HIGH, MEDIUM, and HELD tiers rather than a made-up readiness percentage.

What counts toward the three free mocks?

Only a mock that reaches a completed result counts. Starting, scheduling, abandoning, or failing to finalize a sitting does not use one of the three.

What happens after my third free mock?

Starting another mock requires a paid plan. Your diagnosis, repair plan, completed mock debriefs, and linked evidence remain available. The full repair library, unlimited mocks, live drills, and mastery tracking are included with paid access.

Candidates preparing for rigorous Quant, Machine Learning, AI Engineering, Forward-Deployed Engineering, and Data Science interviews where they must explain and defend their reasoning.

How is the diagnosis graded?

Every published problem has fixed, reviewed criteria. The server records the attempt, the grader must cite what you actually said, and deterministic checks validate those citations and require enough coverage. If the evidence is thin, the result stays inconclusive.

Why not use ChatGPT or a question bank?

Both are useful for practice, but neither gives you this complete loop by default: a fixed reviewed problem, a server-recorded timed attempt, evidence-checked grading, a target-scoped diagnosis, a repair plan, and a proof mock that prioritizes observed weaknesses.

The published curriculum covers probability, combinatorics, statistics, linear algebra, ML foundations, and AI engineering. The separate interview-problem bank also includes standalone questions across finance, brainteasers, applied AI systems, and other technical interview topics.

What happens in a mock interview?

Mocks are text-based. You choose two to four questions plus a role and seniority target. The timer adjusts to the configuration; there are no hints, pauses, or restarts. The debrief links every conclusion to the recorded question session.

What does paid access cost and include?

Basic is $69 per month and can be canceled from billing settings; access continues through the paid period. Lifetime is a one-time $399 payment. Both include the full published repair library, unlimited mocks and live drills, and mastery tracking. Active offers appear automatically at checkout.

Zvsquared — quant interview diagnostic & AI mock interviews

The problem bank includes questions attributed to interviews at

Google
Citadel
Two Sigma
Goldman Sachs
Morgan Stanley
D.E. Shaw
JP Morgan
Amazon
Microsoft
Bloomberg
Stripe
Uber
Airbnb
Netflix
AQR
SIG
Akuna Capital
Optiver
Point72

The problem

You do not need another 300 questions.You need to know which three can still kill your interview.

You may have solved hundreds. You still do not know which answers survive "Why?" or "What did you just assume?"

One diagnostic finds the observed gaps, ranks them, and tells you what to fix first.

The founder

"Why the log?"

Aleksandr Zvonarev — founder of Zvsquared — Aleksandr Zvonarev
Founder
Built Zvsquared after his own WorldQuant interview

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.

How it works

Diagnose. Fix. Prove.

Three free completed mocks form one evidence loop: find the gaps, repair them, then test whether the repair held.

01Diagnose
Find what will break.
One timed diagnostic mock probes every claim and records the criteria your answer did — and did not — defend.
Probability › Conditional probability00:12Saved
Define conditional probability.
DEF · Probability · ~3 min · 2 hints available
Examiner
Define P(A | B). State the condition under which it is defined.
You
P(A | B) = P(A ∩ B) / P(B).
Examiner · Hint1 left · impacts verdict
You gave the formula. What condition must B satisfy, and what does conditioning on B mean?
02Fix
Study only what matters.
Observed weaknesses become a ranked repair plan of targeted drills, so the next study block starts with evidence.
Ranked repair plan
1. 01Conditional expectation3 drills~40 min
2. 02Bayesian updating2 drills~25 min
3. 03Experiment design2 drills~30 min
Illustrative study block — your plan follows the observed risks.
03Prove
Make sure it is actually gone.
The next mock prioritizes weak topics through different reviewed problems where the pool allows. A tier moves only after conclusive evidence.
Fresh evidence
Conditional expectation
HIGHHELD
Fresh problem · no negative criteria

Under the hood, targeted drills move from Define to Understand to Apply. See the fix library.

The diagnostic sitting

Timed. No hints. No restart.

Your first completed mock is your diagnostic — three are free. Choose two to four questions; the time limit adjusts to seniority and question count, and the examiner stays in interview mode throughout.

Find my weak spots — free

00:00

The clock starts immediately. No pause, no restart.

Time limit

A completed sitting ends in one aggregate debrief.

Question 1

Next question

Up to 2 more

Questions unlock one at a time.

The diagnosis

You leave with a ranked list, not a feeling.

The report ranks only weaknesses supported by recorded criteria. It separates technical correctness from interview behavior and says what has not been tested yet.

Example diagnosisQuant Researcher · New Grad

Observed risks

01Conditional expectation2 of 5 criteria negative · independence assumption undefendedHIGH
02Bayesian updating2 of 6 criteria negative · prior choice not justifiedHIGH
03Experiment design1 of 5 criteria negative · stopping rule omittedMEDIUM

Recorded evidence

"You used independence in the final step without showing why the variables were independent."

Not yet tested: stochastic processes · market microstructure · time series

Illustrative report — real reports link each line to the recorded session.

The fix library

Turn each weakness into a drill.

Every weakness in your report maps to drills here — six published pillars plus the standalone interview bank. Each listed problem has fixed, reviewed criteria and can run live.

Reviewed

standalone interview problems

INT

Best-of-Seven Series

Google · Probability

Runs live · evidence-graded

INT

Unfair Coin Detection

Probability fundamentals

Runs live · evidence-graded

INT

Russian Roulette Strategy

Brainteasers

Runs live · evidence-graded

Moreadded after review

INT marks standalone interview problems. Curriculum drills run in three layers: DEF (define), UND (understand), PRB (apply).

The difference

Practice gives you volume. Diagnosis gives you direction.

Where to start

Question banks

Gives you hundreds of questions and leaves the priority to you.

Zvsquared

Ranks the observed weaknesses that need attention first.

After a wrong answer

Question banks

Ends at the answer key or a model solution.

Zvsquared

Ends with recorded evidence and a targeted repair plan.

Reasoning pressure

Question banks

Cannot ask why, challenge an assumption, or change direction.

Zvsquared

Probes each claim until the reasoning holds or a gap becomes visible.

What counts as fixed

Question banks

Completion and confidence become proxies for readiness.

Zvsquared

A risk tier moves only after fresh, conclusive evidence.

What you take away

Question banks

A longer completed list and your own interpretation.

Zvsquared

A diagnosis, repair plan, and proof trail tied to recorded attempts.

	Question banks	Zvsquared
Where to start	Gives you hundreds of questions and leaves the priority to you.	Ranks the observed weaknesses that need attention first.
After a wrong answer	Ends at the answer key or a model solution.	Ends with recorded evidence and a targeted repair plan.
Reasoning pressure	Cannot ask why, challenge an assumption, or change direction.	Probes each claim until the reasoning holds or a gap becomes visible.
What counts as fixed	Completion and confidence become proxies for readiness.	A risk tier moves only after fresh, conclusive evidence.
What you take away	A longer completed list and your own interpretation.	A diagnosis, repair plan, and proof trail tied to recorded attempts.

Why trust the diagnosis?

The result cannot outrun the evidence.

The examiner can vary its wording. The criteria, recorded attempt, and evidence required for a verdict stay fixed.

1
Not generic AI feedback.
Every published problem has fixed, reviewed criteria before it reaches a mock.
2
It grades what you said.
Verdicts must point to evidence from the attempt the server actually recorded.
3
It cannot move the goalposts.
The criteria do not change mid-interview, and thin evidence stays inconclusive.

Pricing

Choose the access that fits your timeline.

Every account starts with a free diagnostic — three completed mocks. Subscribe monthly, or pay once for unlimited mocks, the full repair library, live drills, and mastery tracking.

Pay once

Lifetime

$399one-time

One payment. Full access for the life of the product.

Full published curriculum
Unlimited live sessions + timed mocks
Mastery tracking across sessions
No subscription, no renewals

Get lifetime

Basic

$69/mo

Full access while your subscription is active.

Full published curriculum
Unlimited live sessions + timed mocks
Mastery tracking across sessions
Cancel anytime from billing settings

Choose Basic

Teams & cohorts

Enterprise

Custom

Custom pricing for your team

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

Contact us

Common questions

Asked before buying.

No. Create an account and you can complete three mock interviews for free without adding a payment method.

The real interview won't go easy on you.Neither will this.

Run the free diagnostic. Get your risk map and repair plan before interview day.

Find my weak spots — free Sign in

No card required. Your report stays available.

Know exactly what to fixbefore interview day.

Diagnose. Fix. Prove.

Find what will break.

Study only what matters.

Make sure it is actually gone.

Timed. No hints. No restart.

You leave with a ranked list, not a feeling.

Observed risks

Turn each weakness into a drill.

Best-of-Seven Series

Unfair Coin Detection

Russian Roulette Strategy

Practice gives you volume. Diagnosis gives you direction.

The result cannot outrun the evidence.

Not generic AI feedback.

It grades what you said.

It cannot move the goalposts.

Choose the access that fits your timeline.

Lifetime

Basic

Enterprise

Asked before buying.

The real interview won't go easy on you.Neither will this.