Prepare to explain the decisions behind your project, not just its final chart. A useful discussion connects the question, available information, method, comparison and limitation.
The questions below are original practice prompts. They are not a verified employer question list or a claim about a universal interview format.
Begin with a bounded opening
“I compared a simple forecast with constant baselines in synthetic time-ordered data. The study recovered a planted relationship and showed how the same procedure failed after a sign reversal. I implemented the evaluation and documented the failure; the result is educational research, not trading performance.”
Adapt this structure to your own facts. Do not claim implementation work that came unchanged from a supplied project.
Five questions to practice
What did you personally do? Identify the functions, analyses or decisions you owned and what collaborators or source material supplied.
What was known at prediction time? Pick one row and walk through feature availability and label resolution. A general statement that the data was chronological is insufficient.
Why this baseline? Explain the simpler alternative and which additional claim the candidate model is trying to establish.
What failed? Show an unfavorable setting, invalid initial assumption or unresolved limitation. Distinguish model failure from an implementation defect.
What would you do next? Propose an experiment that separates explanations. Do not describe its hoped-for outcome as an observed result.
A short original dialogue
Reviewer: Why did an expanding refit still miss the reversal?
Candidate: Older positive examples remained most of the training history, so the estimated slope stayed positive. Refitting occurred as specified, but forgetting was slow. A shorter window could adapt faster while increasing estimation noise; this release does not establish the best window.
Reviewer: Why use synthetic data?
Candidate: It lets me know the mechanism and isolate a change. That makes diagnosis easier, but limits external relevance. It does not show that the feature exists in a real market.
Review the explanation
Ask a listener to identify your contribution and the strongest conclusion supported by the evidence. If they infer a broader claim, revise the wording. If they cannot follow the procedure, add the missing contract or comparison before adding more technical terminology.
Use the research memo template to prepare the written version. The Quant Research Project Lab includes the executable study, recorded results and extended discussion prompts.
Read before choosing
Open the actual pages.
7 sample pages, including complete explanations. No email address or account required.
Open the PDF previewPreview page 4 of 7. Use Enlarge page for a closer view. When the page is focused, use left and right arrows to change pages.
