ZVSquared: The AI interviewer that won't accept memorized answers

Hey Product Hunt — I’m Aleksandr, the founder of Zvsquared. I built Zvsquared after failing a WorldQuant interview on a question I thought I understood: logistic regression. The interviewer interrupted me and asked: “Why the log? Why specifically that function?” I had prepared the standard explanation, but I couldn’t defend the reasoning underneath it. That was the moment I realized that knowing an answer and being able to survive follow-up questions are completely different skills. Most interview preparation rewards recognition, polished responses, or getting the final answer right. Strong technical interviewers keep probing until they find the assumption you cannot justify. Zvsquared is the examiner I wish I had. It challenges every step of your reasoning, asks follow-up questions when an explanation is weak, and evaluates your answers against locked, problem-specific rubrics. Instead of only telling you that an answer was wrong, it records the exact criterion you failed to defend so you know what to work on next. At launch, Zvsquared includes: — 319 hand-reviewed, rubric-graded problems — Probability, combinatorics, statistics, and linear algebra curricula — Additional interview problems in ML, finance, coding, and brainteasers — Timed mock interviews with no hints, pauses, or restarts — Saved transcripts, verdicts, missed criteria, and mastery history This is deliberately not a friendly chatbot that quickly reveals the solution. It is designed to expose gaps before the real interviewer does. I’d especially value feedback from people preparing for quant, ML, data science, and research interviews: What would make the examiner feel more realistic—or more brutal?