The preparation is in the reasoning.
Practice quant interview problems, complete a reproducible Python lab, and improve how you present your technical experience.
Probability. Research. Evidence.
Three books. One standard: show your working.
Choose a box with these probabilities, then draw one token uniformly from it. You see the color, not the box.
A contributes ¼ probability of blue; B contributes ⅙. Among all blue outcomes, A’s share is (¼) ÷ (¼ + ⅙) = ⅗.
Same tokens. Different prior. Different answer.
Choose a starting point
Each path starts with free material. Follow the steps, then decide whether you need a book or written feedback.
THE RESEARCH STRESS LAB
A feature arrives late. A holdout becomes part of the search. Costs erase the margin. Inspect three synthetic studies and write the limitation that belongs in the memo.
Open the free research labThree cases with explanations. No Python setup required.
The working library
Read the sample. Inspect the method.
Choose the preparation you actually need.
01 / Probability & trading
Conditioning, expectation, stopping and uncertainty. Forty original exercises, full solutions and two mock sessions that change the assumptions.
02 / Research projects
A reproducible synthetic study, from the data contract to the research memo. Investigate leakage, changing relationships and rolling estimates.
03 / Career preparation
Choose a role direction, translate technical work and organize the next six weeks. Three worked candidate cases and editable planning sheets.
Working across all three areas?
Compare the complete libraryThe resume is an argument.
A written review connects your experience to the evidence a reader needs: what you did, how you evaluated it, and what the result means.
Read a complete sample reviewFictional teaching example. No hiring outcome is implied.
An evidence edit / Synthetic project
The original claim
“Built a profitable quant model.”
A more specific account
“Compared a frozen linear baseline with rolling estimates on synthetic time-series data; documented where a changed relationship made the original fit fail.”
Use this wording only if the underlying work supports it. State the dataset, comparison and failure—not an unsupported trading outcome.
Choose an initial role direction by responsibilities and evidence, then check the actual listing instead of relying on the title alone.
02 / Examine an assumptionWork through an original detector example, distinguish sensitivity from a posterior, and check the effect of a rarer event.
03 / Explain a resultPresent the question, procedure, result, counterevidence and next experiment in a compact memo that connects to reproducible files.
04 / Show your contributionTurn a vague project claim into a defensible bullet using the question, personal contribution, evaluation and limits.
A model before a number
Work through the original diagnostic, compare your reasoning with full solutions, then choose one thing to repair.
Open the free diagnostic