I get asked about this often enough that it’s worth writing down. This isn’t a universal formula, just what I’d tell a past version of myself.
1. Relax
Applying can be super stressful, and that’s normal. Pace yourself, keep one calendar with every deadline on it, and don’t measure your worth by a single cycle.
2. Make sure you actually want this
A PhD is five-plus years of research. Before anything else, check that research is something you genuinely enjoy and can see a career in, whether that’s academia, an industry lab, or somewhere in between.
3. Make two lists
First, the research topics you want to work on. Second, the professors who are actually taking students this cycle. Departing, tenure-track and lab-size signals matter; email current students if you’re unsure.
4. Apply where they overlap most
Prioritize the professors whose recent work hits the most topics on your list. Committees admit to a lab, not to a department, so advisor fit matters more than program prestige.
5. Read before you reach out
Skim 3–5 recent papers from each professor (the method and discussion, not just the abstract), then send a short email: who you are, one specific thing about their work, and whether they’re taking students. Don’t attach a CV unless asked.
6. Write a focused SOP and personal statement
Your statement of purpose is a fit document, not a highlight reel. A statement that works does three things:
- States a real, specific research question you care about, not “I’m interested in AI.”
- Shows evidence you can do research: a project, however small, with a method and a result.
- Connects that question to why this specific lab, citing actual papers and methods.
If a program also asks for a personal statement, use it for the story behind your path: what drew you to research and what shaped you. Tailor both for every program.
7. Get letters from professors you’ve actually worked with
Letters carry a lot of weight, and the strongest come from professors who supervised your research and can speak to how you think, not from a famous name who only knows your grade. Ask at least four to six weeks ahead, and send each writer a one-page brief: your CV, the programs and deadlines, and two or three specific things you’d love them to mention.
8. Mechanics people forget
- Contact current students, not just the advisor. They’ll tell you about funding, advising style and lab culture.
- Track everything in one spreadsheet: deadlines, portal links, letter status, advisor contact.
- Interviews are mutual. Come with real questions about lab structure, funding, and what a typical week looks like.
9. What I’d underweight
GRE-equivalent scores, long lists of unrelated extracurriculars, and trying to sound like you already know exactly what your dissertation will be. Committees are checking for research taste and trajectory, not a finished plan.