The grad-school resume is not your job resume

Your reader this time is a professor

HR scans for whether you can start work now, seven seconds at a time. A committee reads longer and hunts for one signal: research potential. Can you define a question, choose a method, and finish something that has no answer key?

So the trophies of a job resume (clients, revenue, business impact) fade into the background here. Two weeks spent fixing a dataset nobody would give you matter more than a quarter of sales quota.

The order flips: education on top

• Education comes first, packed with detail. Give the degree, the department, and the GPA with its scale (3.8/4.0, never make them guess). Add rank if strong, plus four to six courses related to the target field.
• Give awards a base number: “Dean’s List, top 3% of 220” means something. A prize name alone does not.
• Work experience moves down and gets cut to what supports the application. On a CS application, your marketing internship is one line, or none.
• One page. The three-page “CV” is for professor jobs. For a master’s, dense beats long.

Coursework is six courses of evidence, not a copy of your transcript

The department already receives your transcript. Retyping thirty course titles burns half a page and tells the committee nothing it did not have an hour ago. This section has one narrow job: show that what you have already taken can carry their first-year requirements.

That means you read their curriculum before you pick. Open the program page, find the required courses for year one, then go back through your own transcript and pull the four to six that line up with them. A course that does not line up stays off, however good the grade was.

• Quantitative training goes first: statistics, linear algebra, algorithms, econometrics, experimental design. The committee is not worried that you are not clever. It is worried the math drowns you in the first semester.
• Mark any graduate-level course you took as an undergraduate and say so in the line. It is usually the strongest line in the section.
• Use the official English title from your transcript, not your own translation. When the course on your resume does not match the course on file, someone has to stop and check that you are not inflating.
• Give grades for all of them or for none. Listing grades on some courses points the committee straight at the ones you were hiding.
• Leave MOOCs and self-study out of this section. If a course actually changed your work, it belongs in research experience as a method: you used the tool to do the thing. That reads far better than a completion certificate.

Writing research experience without exaggerating

The common panic: “I have no publications, so I have no research experience.” Committees do not expect publications from master’s applicants. They expect the shape of research: a question, a method, your own contribution, an outcome.

Course projects, a senior thesis, RA work, even a careful analysis at your job all count. Write each one the same way: what was the question, what did you try, which part was yours (not “we”), and what came out.

• Weak: “Participated in a machine learning research project.”
• Strong: “Built the data pipeline for a 40k-image classification study; ran the ablation experiments that identified the two features driving most of the accuracy.”
• A negative result is still a result. “The result led the lab to drop its original approach” is a legitimate line, and often a good one.

Name the professor and the lab, because that is your coordinate

Academia is smaller than you think. A line reading “Research Assistant, Institute of Information Science” gives the committee an anonymous unit. A line reading “Research Assistant to Prof. [name], [lab name]” gives it a name somebody at that table may already know, and the credibility of the whole entry changes.

For the same reason, the lab’s direction and the keywords of your problem matter more than your title. The committee is placing you on a map: which subfield, near whom, on what class of question. “Research Assistant” by itself carries no information at all.

• Spell out collaborations across labs, departments, or countries, and name the other side. Naming the group and the school is what makes the line checkable. “Participated in an international project” is not.
• Research done inside a company counts the same way. Name whoever led it, PI or senior researcher, and ask them before you put their name on a document.
• If your supervisor prefers not to be named, or the work is unpublished and the topic is confidential, write the field and the method and leave the topic out. This comes up far less often than applicants fear.
• Three RA stints of two or three months each read as someone who does not stay. One entry written to full depth beats three shallow ones.

The publications section has to be graded honestly

Having no publications as a master’s applicant is normal and costs you nothing. What costs you is writing “submitted, still under review” so that it looks published. Some of the people reading your file review for those journals. They see it at a glance, and once they do they go back and reread every other line you wrote.

So label the stage each item has actually reached, and leave no room to guess.

• Published: full citation, authors in the original order. Do not move yourself to the front.
• Accepted but not yet out: add “accepted” or “in press”, with the journal name.
• Under review: write “under review” and the journal. A journal name with no status attached is exactly the blur described above.
• In preparation: you may list one. Listing three tells the committee you have never finished anything.
• Conference posters, workshop papers, and undergraduate symposia all count, but give them their own heading. Do not pad the journal list with them.
• For a preprint, give the arXiv or SSRN link. The link is the evidence. Adjectives are not.
• If you have none of this, delete the heading. An empty section is worse than no section, and those two lines are better spent on research experience.

Numbers still rule, but the kinds change

Job resumes count revenue. Academic ones count sample sizes, dataset size, number of experiments, rankings, scholarship amounts, and students taught. The rule is the same either way: a claim without a number reads as just an adjective.

In the editor, switch the industry style to Grad School. The page then follows these rules: education first, research as evidence, verbs like Researched, Analyzed, Co-authored. Missing numbers still become [N] blanks, and the admissions interview will question your figures like any other interview.

The master’s version and the PhD version are read by different people

Master’s applications usually go to one committee working through a large pile quickly. It is deciding whether you can survive two years of coursework and a thesis. Full coursework, a clear GPA, and research with a recognizable shape are enough, and one page is the right length.

A PhD application is really read by the two or three professors who might take you into a lab. They are deciding something harder: whether you can run for four to six years, including a whole year with no progress, without falling apart. So the research entries have to reach the level of method, meaning how you designed it, where you got stuck, and how you got around it. Two pages is reasonable there.

• A PhD version can carry three lines of research interests. Write directions and methods, not one locked topic. Too narrow, and the professor who does not work on that exact question skips you.
• Make the skills line concrete on the PhD version: languages, packages, instruments, databases, lab techniques. That is what a professor uses to estimate how long you will take to train.
• The master’s version needs the opposite restraint. Stretching coursework and research to two pages hides the point from someone reading forty files today.
• Applying to both in the same season means writing both. The first section alone is ordered differently.

The resume has to hold up your SOP and feed your recommenders

The statement of purpose is the argument: why this field, why this program, why now. The resume is the index: where each claim happened, at what scale, and when. Keep that division clean and the two documents stop undercutting each other.

Reconcile them before you submit. Take every factual sentence in the SOP and find the line on the resume that backs it. If the statement says two years in the lab and the resume shows three months, the committee catches it, and then it rereads the whole file looking for the next one.

Check the other direction too. Does the SOP open up your strongest entry? If your best research is never mentioned in the statement, you have demoted it to background yourself.

• Do not write SOP sentences on the resume. “I am passionate about this field” wastes a line. The resume holds only what can be checked.
• Most professors ask for your resume and a draft SOP before they write anything, because their memory of you is less detailed than you assume. The version you hand over decides which details reach the letter.
• So the recommender’s copy can be thicker than the submitted one. Expand the work you did with that person into three or four lines: dates, your part, what came out. Give them something to quote.
• Split the three letters: one on research, one on the classroom, one on somewhere else. When you ask, point at the entry you hope that person will cover. It works better than a long polite email.
• Do not list referees or their contact details. Academic applications collect that inside the system, and “References available upon request” is a wasted line.

If you are going back to school after years of work

Applicants with a career behind them make two opposite mistakes. One sends the job resume untouched: five roles, no research, nothing the committee can read as research potential. The other loses their nerve, deletes the career, and submits a page whose last entry is an undergraduate project from four years ago, followed by a blank.

The move is translation, not deletion. Your job already contained research: a market analysis, an A/B test, a model you put into production, a process improvement, a quality investigation. Anything with a hypothesis, a method, data, and a conclusion can be rewritten in the language of research.

• Take the one or two roles that look most like research and expand them to four or five lines. Compress the rest to one line each. Senior Manager earns you nothing here. The analysis you designed and ran yourself does.
• Drop the corporate vocabulary. Stakeholder, deliverable, and roadmap are noise on an academic resume. Hypothesis, method, sample, and limitation are not.
• If you have been away from school a long time, one recent course or a part-time RA stint is the most valuable line on the page. It proves not knowledge but that you can get back into that mode.
• Do not leave a hole in the timeline. A year spent on exams, on military service, or caring for a family member is one line. The blank looks worse than the explanation.

Ten minutes before you submit

• Name the file Lastname_Firstname_CV.pdf. A department office downloads hundreds of files a day, and cv_final_v3.pdf is missing the moment it lands in the folder.
• PDF only, and open the PDF yourself after exporting. Fonts that failed to embed and links that flattened into plain text both happen at the moment of export, not before.
• No photo, no date of birth, no gender, no national ID number, no marital status. North American applications do not ask for any of it, and volunteering it only gives the reader something to pretend not to have seen.
• Spell out every abbreviation the first time it appears. University names, prizes, institutes, and lab short forms read as meaningless strings of letters outside your own country.
• One date format all the way down. Sep 2024 – Jun 2025 is fine. Do not switch to numbers halfway through.
• Change two things for every school: the six courses, and the order of the research entries. A file that is identical everywhere reads like a backup plan everywhere.
• Print it, or at least zoom out to a full page. One page means one page of paper, not one scroll of a screen.