What CS Professors Look For in an SOP

What CS Professors Say They Look For in an SOP

Computer science is unusually well documented on this question. Faculty publish long applicant guides on their own department web pages and leave them up for a decade. Departments post the criteria their own faculty compiled. You do not have to infer what CS readers want; several of them wrote it down and signed it.

Below are eight public sources — four named faculty pages and four department pages — quoted from their live text.

1. Your reader is a computer scientist, not an admissions officer

Harvey Mudd College's computer science department opens its guide with the single most consequential fact about the document:

"Keep in mind that your essay (and all of your application materials), will be read by professors of computer science and not by admissions officers. You should assume that your readers are very well-trained in computer science."

Almost every other piece of CS-specific advice follows from that sentence. You are not explaining your field to a generalist. You are not selling. You are writing a short technical document to someone who already knows the area and can check your claims.

2. They are reading it as a research statement

Mor Harchol-Balter, the Bruce J. Nelson Professor of Computer Science at Carnegie Mellon, says the label is misleading in her long-standing applicant guide:

It's misleading that the personal statement is called a "personal" statement, since what admission committees are really looking for is a research statement. What admission committees want is a statement about what research you have done, what research you hope to do, and why you like research.

Harvey Mudd's page lists what its readers want conveyed, and the list is the same shape: that you are genuinely interested in research, that you can speak about a subdiscipline "with some level of sophistication", that you have had research experience, that you "understand the research process from selecting a problem to solving it to presenting your results", and that you have done your homework on the specific department.

3. The document has a routing job

Carnegie Mellon's Jonathan Aldrich — Professor in the Software and Societal Systems Department and the Computer Science Department — names the mechanism that most applicants never see:

"Mention the names of specific faculty if you know you want to work with them. This helps to ensure the right faculty will read and evaluate your application. It's OK to mention more than one name, but naming the whole department is superfluous."

Norman Ramsey, now a retired professor (emeritus) at Tufts University, describes the same function from the committee's side, and attaches the condition:

"The admissions committee will use your statement of purpose to try to figure out who might be your research supervisor. If you know you want to work with particular faculty, by all means mention them by name. Just be sure to demonstrate you know something about them--if you appear to have picked a name out of a hat, you are unlikely to be admitted."

The faculty who compiled CU Boulder's PhD statement guidance ask for the same specificity: "What research area(s) you are interested in and, specifically, which professors at CU Boulder would be a good potential fit. Consult the faculty's research to find specific overlaps of your interests with ours."

The University of Toronto's computer science department says why it matters: "Mention the professors with whom you would want to work in the department that you are applying to, and why you think your interests align. This shows that you've done your homework on the department."

4. What to cut

This is where CS advice is bluntest, and the blame lands on one specific opening.

Andy Pavlo, an associate professor in the Computer Science Department at Carnegie Mellon, wrote a whole post about the statement that fails:

"Professors don't care that you got the first computer on your street and started messing with it when you were really young. We don't believe you either."

Toronto's department reaches the same conclusion in capital letters: "Do NOT include tales about your childhood and how you've always loved computers. This type of information is irrelevant and suggests that you are unfocused or have little to say."

Aldrich's page permits a version of the same material — but bounded. He asks you to say why research in your field matters to you, and if your background is unusual, to explain how it bears on your PhD plans, then adds: "This section can provide useful context, but keep it relatively short; spend most of your time on the research you've done and what you hope to do in the future."

CU Boulder's faculty guidance draws the line in one sentence: "Apart from the above, an overly-lengthy personal background is generally not helpful."

5. What to put in instead

Aldrich's page is effectively a spec, and it is worth reading as one. He asks applicants to describe research they have done — specifically, that you can "describe the problem you confronted, argue that it is important and was challenging to solve, compare (briefly) to prior work, and explain your technical solution and how you validated it at a high-level." Then he asks for a forward problem:

"Tell us about some problem you are interested in. It doesn't have to be the only problem you want to work on, but it should be a good scientific problem. You should explain what is the scientific question, why it is important, why it is hard, and why it is open."

Harchol-Balter's guide supplies a paragraph-by-paragraph template with the same content in a different order: areas of interest first (so the committee can route your file to the right readers), then several paragraphs on specific research projects — "What was the problem you were trying to solve? Why was it important? What approaches did you try? What did you learn?" — then why you need a PhD, then why this department.

Ramsey adds a smaller point that is easy to act on: "Another good thing you can do in a statement of purpose is give clear, concise descriptions of work you have done. This is an important skill for a researcher, and you have an opportunity to show that you are already developing this skill. For bonus points, explain the significance of your work."

CU Boulder's page asks for the same granularity about your own contribution — "Be specific about your role and what you found exciting. Consider including some technical details." — and then widens what counts as relevant experience: "We recognize that not all students were able to access such opportunities. We welcome other work or personal experiences that, even if not directly relevant, demonstrate your experience in overcoming challenges or working with others."

6. Naming an area is not a binding commitment

Two of the strongest research-first voices in CS also supply the reassurance.

Harchol-Balter: "It's important to realize that the research statement is not a commitment to do research in that area. A third of all applicants end up working in an area different from that which they described on their research statement."

Harvey Mudd, on naming a subdiscipline in the first paragraph: "This is not a contract that forces you to study that subdiscipline and you might change your mind later."

Aldrich says the same thing about the forward problem: "It doesn't have to be the only problem you want to work on, but it should be a good scientific problem."

The instruction to be specific and the permission to change your mind are not in tension. Specificity is evidence that you understand the field well enough to pick a problem. It is not a signature.

7. Length: one to two pages, and the department will say so

Carnegie Mellon's School of Computer Science asks for "a concise one- or two-page essay that describes your research interests, related experiences and objectives in pursuing a graduate degree at Carnegie Mellon."

CU Boulder is explicit that it sets no rule and then sets one anyway: "There are no word or page limits or any format for this SOP. However, keeping it concise to a page or two will be appreciated."

Toronto handles the silent case: "If a page limit is not specified, then two pages is generally a good length for a Statement of Purpose. Less than two pages could suggest to the committee that you don't have enough to say, while more than two pages could mean that your application materials have not been well edited."

CU Boulder also suggests a formatting move that few applicants use: "Consider using small section headers to signpost the important parts of your statement, such as area(s) of interest and research experience."

Where CS readers still disagree

Even inside one field, the guidance is not unanimous, and the differences are worth knowing before you write.

How central the person is. Pavlo and Toronto would delete your origin story outright. Aldrich keeps a bounded version of it up front. CU Boulder invites experiences "even if not directly relevant" as evidence of persistence and collaboration. Same field, three different tolerances.

Where fit goes. Harvey Mudd puts the department-specific paragraph last. Harchol-Balter puts areas of interest first and CMU-specific fit last. CU Boulder wants the faculty-fit material in the very first component. If the program publishes a structure, use theirs.

How much technical detail. CU Boulder explicitly invites technical details. Harvey Mudd assumes readers "very well-trained in computer science." Aldrich asks for the technical solution "at a high-level." Calibrate to the page you are reading, not to a general rule.

What none of them disagree about: the statement is read by a researcher, it is judged on evidence, and vague enthusiasm is a cost, not a benefit. If you want to know whether your draft clears that bar, a review against a published rubric will show you which of these expectations it already meets.

Related reading

• What professors across disciplines say about the SOP — the cross-field hub this CS-specific page extends

• Should you name a professor in your SOP? — the naming question across fields

• Research proposal or life story? — the same argument outside CS

• SOP length: 500 words vs 3 pages — why length advice conflicts

• MS SOP: research track vs coursework track — what changes for a taught master's

• Technical language in a statement of purpose — how specific to get

All quotations in this article were re-checked against their live source pages on 16 August 2026. Faculty pages and departmental guidance change; verify against the program page you are applying to.