Resources

Interview guide

How to prepare for a PhD advisor interview, what to ask, and the traps that hurt even strong applicants.

Before the call (48 hours out)

Re-read the professor's last two papers top to bottom — not just abstracts. Mark the sections that genuinely surprised you.
Check the lab website for who recently graduated and where they went. This tells you more about trajectory than any rankings do.
Open the PhdFit meeting-prep sheet — the Meeting Prep Compiler pulls these signals automatically.
Write down three questions that only make sense if you read the papers carefully. Those are the ones worth asking.

Lead with a specific thread from their work. "I was struck by your handling of X in the 2024 paper" beats "I'm interested in your research."
Be honest about your background. Overstating experience falls apart within 60 seconds of a real technical question.
Let the professor drive. They will tell you which project they care about most — that's the project to talk about.

"How do you and your students decide what to work on?" — reveals whether the lab is directive or self-driven.
"What has surprised you about the research agenda in the last year?" — opens a window into how the professor updates their views.
"What do your students who graduate into industry research wish they'd started earlier?" — produces concrete trajectory advice.
"Are you admitting students this cycle?" — ask plainly near the end, not between the lines. Polite directness is read as maturity.

Asking questions already answered on the lab website. Trains the professor to see you as under-prepared.
Pitching a specific project you want to do. Most PIs want to see how you think, not whether you already have a plan. Save that for after admission.
Treating the call as a test. It's mutual. You should leave with a read on whether this lab is somewhere you'd survive three difficult years.

Send a thank-you within 24 hours — specific, short, references one thing you discussed.
Write yourself a private note: energy, fit, red flags. You will have forgotten the texture of the call in a week.
Update the PhdFit shortlist bucket (reach / target / safer / not sure) based on what you learned, not what you hoped.