Before preparing for a startup's interview loop, how do you find out which rounds it actually contains?

## Why this is a real skill and not just admin In a standardised loop the stages are published, so preparation is a syllabus problem. In a startup loop the stages are a company-specific decision, sometimes made the week you applied. If you do not ask, you will prepare from a mental template built out of other people's interview stories, and that template is usually a big-company one: automated assessment, algorithm screens, a design round, a behavioural round. A three-conversation seed-stage loop contains none of those, and the hours you spent are gone. ## The four questions that convert an unknown loop into a known one Send them as one short, low-friction message, before or right after the first scheduling email: 1. **What are the conversations, and in what order?** You are asking for the stage list, not a promise. 2. **Who is in each one?** A founder, an engineer you would work with, a non-engineering co-founder and a prospective peer all evaluate different things and can answer different questions of yours. 3. **Is there a practical component, and what shape?** A scoped take-home, a paired session on live code, or a longer trial engagement are three very different asks on your time. 4. **What are you deciding at the end?** This surfaces whether there is a written rubric, an advisory team round, or one person's judgment. A fifth, when the loop is long or you have competing timelines: **what is the expected end-to-end duration?** At seed stage the honest answer is usually one to two weeks. ## The artefact to ask for by name Many small companies write a one-page note for candidates: the stage list, roughly what each round involves, sometimes the trial-project brief attached. Asking "do you have a short write-up of how you hire that I could read first?" is normal, costs the founder nothing, and gets you a document rather than a paraphrase. Two useful side effects: a company that has one has thought about fairness, and a company that cannot describe its own loop has told you something worth knowing. ## Reading the map correctly The answers change what you do with your preparation hours: | What the map says | What to spend the hours on | |---|---| | Founder chat plus paired session on live code | The product, its domain, how you would scope work in it; get fluent reading unfamiliar code quickly | | Take-home plus a walkthrough conversation | Scoping, finishing, and being able to defend your choices out loud | | Trial engagement of a few days | Terms and logistics first, then the work itself | | Team round with the two existing engineers | Concrete examples of how you work with others, and questions of your own | | A design or architecture conversation added | Structured practice at explaining tradeoffs at the scale that company actually operates | ## A worked example A candidate is talking to a nine-person seed-stage company about a full-stack role. They ask for the *how we hire* note before scheduling. It comes back as one page: a forty-minute founder call, a paired session on a real bug from the product queue, a half-hour with the two engineers on the team, eleven days from first call to decision, founder decides. Nothing on that page is an algorithm round. The candidate reallocates: two evenings reading the public product and writing down what they do not understand about the domain, one evening practising reading unfamiliar code under time pressure, and a list of questions for the founder about ownership and roadmap. Their friend, interviewing at a similar company the same month, spent three weeks on puzzle drills instead and walked into a paired session cold. Same market, same seniority, opposite return on the hours. ## When the map turns out to be wrong Unstandardised loops flex, and a stage may be added mid-process. Treat that as normal, and re-ask: "happy to do that; what will that conversation cover and who is in it?" It becomes a genuine concern only when the additions keep growing the unpaid work, or when nobody will say what the added round is for. ## The failure this prevents The single most common wasted preparation in small-company hiring is drilling algorithm puzzles for a loop that never asks one. It feels like diligent work because it is measurable and gamified, and it is the wrong syllabus. The four questions above cost one message and remove the guess.