What a question template is
A scored template contains a category, difficulty, prompt generator, answer function and worked explanation. The generator changes inputs such as probabilities, payoffs, sequence rules, sample sizes, strikes, rates or market conditions.
The current bank contains 222 reasoning families across 16 categories, plus 72 self-evaluated verbal rehearsal prompts. These are tested generators and structured prompts, not a marketing estimate of every possible surface variant.
Evidence reviewed for V16
Reviewed 8 September 2026. Employer sources establish preparation topics, not exact generated questions. Our new exercises are original implementations; role relevance is an editorial inference.
- Jane Street trading interviews: collaborative problem solving, probability, statistics, coding and data analysis. No finance knowledge prerequisite is implied.
- Two Sigma research interview guidance: open-ended data analysis, coding, algorithms and statistics or research-domain discussion.
- Citadel campus research process (9 November 2024): practical problem solving, programming and research discussion; rechecked for this release.
- HRT engineering interview guidance (2 October 2025): engineering scope. Numeric algorithm drills are not a replacement for writing and testing code.
- Five Rings positions: research, mathematical communication and implementation skills. This is role evidence, not a verified interview sequence.
Learning evidence and limits
Roediger and Karpicke (2006) supports retrieval practice for delayed retention. Rohrer and Taylor (2007) studied distributed and mixed mathematics practice. QuantPrep applies these principles through fresh-variant retrieval and expanding intervals. The particular intervals and mastery weights are product heuristics, not clinically validated or experimentally proven optimal for quant interviews.
How the count changed
The earlier headline counted registered generators, including numbered difficulty and range variants. V16 groups these into named reasoning families and maps six overlapping additions onto existing families. Legacy identities remain intact to preserve old reviews. Structural duplicate candidates are recorded for further mathematical review; a family count is not proof that every family is semantically unrelated.
How topics are selected
We use official interview guidance where available, current role material, first-hand public reports and established mathematical interview archetypes. Official sources receive the highest weight. Candidate reports are treated as evidence of recurring themes, not proof that an exact question remains in use.
Source hierarchy
- Official employer interview and careers guidance.
- Role descriptions and training-program descriptions.
- Multiple independent first-hand public reports.
- Long-standing mathematical, research and trading interview archetypes.
What topic evidence means
Topic metadata means a question matches a publicly documented skill or recurring format. It does not mean:
- the exact generated wording was asked by an employer;
- an employer supplied or endorsed the question;
- a current interview will contain that family;
- one preparation path covers every role, desk or level.
Originality and proprietary material
Reported questions are abstracted into a mathematical family and rewritten as original generators. The product teaches methods rather than distributing private or proprietary assessment content.
Quality-control process
The latest bank expansion executes every generator repeatedly and checks for a non-empty prompt, finite numeric answer, worked explanation, registered category and unique ID. Coverage gates also require the newly audited confidence, modeling, algorithm, quoting, hedging, estimation and information-reasoning families.
The build runs 120,832 seeded generator checks, plus structural validation of all 72 verbal prompts. This reduces implementation errors but does not imply that every mathematical explanation can never be improved. Version notes document major corrections and expansions.
Known limitations
- Interview processes change and differ by role, office, desk and level.
- Research and developer roles require hands-on coding and deep discussion of the candidate's own work beyond a general trainer.
- Systems and low-level engineering interviews are outside the core quantitative-trader scope.
- No question database can predict every future prompt or guarantee an offer.
Independence
QuantPrep is an independent educational project. It is not affiliated with or endorsed by Jane Street, Optiver, IMC, Susquehanna, Citadel Securities or any other employer named on the site.
Turn the guide into timed practice
Select a category and difficulty, then run a short session and review every miss.
Start a fresh session