Which questions are worth asking the firm at the end of an interview? This note lists a few that match the public standard, using the boundary around a research notebook and a live order as the example, for candidates in New York, London, and Hong Kong.
Ask about the boundary and the review
A worthwhile question reveals how the work is actually governed. Ask who reviews a change before it leaves a notebook. Ask what happens to a model that fails its tests. Ask whether an internship ships a tool or a tour. Ask how New York, London, and Hong Kong share one process rather than three. These questions have answers a serious interviewer can give without inventing a result, a headcount, or a speed figure. If the answers are vague, the vagueness is data. If the answers contradict the public notes, that contradiction is also data.
Questions about pay structure in general, location, and the shape of the interview itself are fair. Questions that ask the interviewer to disclose positions, promise a return, or describe a personal colleague’s private life are not. The second kind puts the interviewer in a hole. A good process will decline them. A candidate should not read the decline as hostility. Personal information is not an interview souvenir.
A question worth asking the firm is one that tests a published boundary: review, the removal of failed work, the limit of an internship, or the wall between a notebook and an order.
One question that tests the wall
A specific question worth asking is what would happen if a notebook studying a public name, say SPY, were pointed at a host that can send orders. The answer consistent with the firm’s public engineering and policy notes is that this would be a broken boundary, not a sign of trust. Interns are not the exception. Experienced hires are not the exception. An interviewer who answers that the boundary depends on how promising the notebook looks has told the candidate something important, and it is not comforting. The SPY notebook is only the example. Any public instrument would serve.
Other versions of the same probe ask how a dataset is accepted, how a day is logged so it can be replayed, and how a person in Hong Kong reviews work that a person in New York will rely on while the U.S. session is open. The point of mentioning all three offices is that the answer should not be that one office is where the real work happens. New York, London, and Hong Kong are the offices the site already lists. A recruiting answer that invents a fourth center of gravity, or that waves the other two away, is not describing the firm that published the site.
Internship questions belong in the same family. What does a Friday deliverable look like? Who reviews it? What is explicitly out of scope? A summer that cannot answer those questions in the interview will not become clear in July. Candidates should ask before they move. The firm should be grateful for the question. An interview loop that treats the question as naive is a loop that wanted an audience, not a colleague.
Questions that are not worth the time
It is not worth asking the firm to confirm a rumor, to rank itself against a competitor by a performance figure, or to translate the public educational pages into a product pitch. Those questions invite the firm to abandon the standard the pages set. A candidate who wants a product pitch is in the wrong conversation. A candidate who wants to know whether the standard is real should ask about review, removal of failed work, and the wall between research and orders.
Questions worth asking are questions the public notes have already committed the firm to answer. The notebook and the order path are the example. Internships, interviews, and the three offices are the context. A process that cannot discuss them is a process the candidate has now measured.
This piece does not script an interview, promise answers, or list openings. It suggests questions that match the public standard, for candidates considering internships or experienced roles in New York, London, and Hong Kong. A notebook pointed at SPY is the example of the wall worth asking about.
