What this is — and what it isn't

How it works

Interview Station runs timed, spoken practice interviews for Australian medical school admissions — MMI stations and panel interviews — and specialty college selection. You read a prompt, the clock starts, and an AI interviewer asks questions, listens to your spoken answers, and follows up on what you actually said. Panel formats run two interviewer voices; MMI role-play stations put an acted counterpart in the room. Afterwards you get a written report: an overall rating, comments by domain, and scores on communication, reasoning, empathy, integrity and insight. The interviewer is built on a large language model, with interview conduct designed by an Australian consultant-level doctor who has sat these selection processes as a candidate and run them as an interviewer.

Why it exists

Candidates consistently report that the real interview felt harder than anything they practised — not because interviewers are hostile, but because a good panel never lets a rehearsed answer stand. They ask why, ask for the specific example, and put the other side of your argument to you. Most preparation is written and solitary, so that experience arrives for the first time on the day. This tool exists to make being probed out loud a routine you can repeat as often as you need, at any hour, with nobody watching.

What it deliberately doesn't do

It does not use any university's, medical school's, or specialty college's actual interview questions, scenarios, or marking documents — no past questions, no remembered questions, nothing recalled from real circuits. Every scenario is an AI-generated original about the publicly known anatomy of these interviews: the same station types and question domains any preparation book describes. It also makes no claims about any named institution's format — formats vary and change, and the institutions are the only authority on their own processes.

It does not promise outcomes. Practice improves fluency under questioning; nothing guarantees an offer, and no score here predicts your result. It is not affiliated with, endorsed by, or connected to any university or college, and it is not an admissions consultancy — nobody here will write your application or tell you what to say.

How to use it well

Interview Station supplements human practice — it does not replace it. Mock interviews with people who will look you in the eye remain the gold standard: they read your face, and they give you the irreplaceable experience of performing under a real gaze.

What an AI interviewer adds is volume and availability. A mock with a mentor takes organising; this takes ninety seconds, at 6am or midnight, and stumbling costs you nothing. Start early — being questioned shows you where your thinking is thin while there is still time to fix it — and use the written reports to see the patterns across stations, not to relitigate single marks. Then take what you've found back to human practice and pressure-test it.

The AI can be wrong — treat it accordingly

Large language models are capable but imperfect. The interviewer can misread an answer, mark unfairly on occasion, and state things with more confidence than they deserve. Speech recognition can also mishear you. Treat the feedback as one experienced opinion, not a ruling: where it points at a pattern you half-knew was there, act on it; where a single mark seems wrong, it might be. Nothing here is advice about what any real institution wants.