JACQUELINE R. M. A. MAASCH

Welcome to the hidden curriculum.

Advice on succeeding as a PhD student.

Note. This page is actively under construction.

Preface.

I began thinking explicitly about the hidden curriculum when I co-designed the first-year seminar for Cornell's computer science PhD program. This page contains my reflections as I reach graduation. The opinions here are informed by my personal journey in the PhD and my second-hand observations on the experiences of my peers. My views have also been shaped by my time as a PhD application reviewer (2 years), PhD recruitment leader (3 years), and peer mentor (1 year) in Cornell's computer science department.

Remark 1. Some of the opinions I share will be relevant across disciplines. It is likely that most of this advice is local to STEM PhDs, and specifically to computing PhDs. My experiences are unique. I can only speak from the vantage point of my own demographic (Millennial, white, not a man, from New England; three of my elders are professors). Take everything with a grain of salt.

Remark 2. The advice here does not indulge much in the romance of science, academia, and research. However, I would hope for all of us that we feel genuine passion for learning and for pushing the boundaries of human knowledge. My love of learning and creating is perhaps my greatest source of joy and gratification. That said, it is important to approach graduate school with pragmatism. Things aren't casual anymore. There are serious tangible consequences to your decision-making. This post focuses on practical aspects.

Some definitions.

The hidden curriculum of graduate school. This is the insider information that can significantly contribute to your success as a doctoral student, yet is rarely explicitly discussed.

The hidden curriculum. The collection of unspoken, unofficial, and sometimes unintended lessons, values, and norms of the field that are learned in graduate school, as distinct from the formal academic curriculum.

Early access to the hidden curriculum is just one reason why people who grew up in proximity to PhDs often seem like they are "ahead of the game." These implicit keys-to-success can have real material impacts on your job placements and financial outcomes. They contribute to your social capital and cultural capital in the field. They can even have a technical impact on your science.

Eustress versus distress. The PhD will be stressful. Ideally, most of your stress will be eustress. Each of us experiences a unique proportion of distress as well. I think of eustress as Type I or Type II fun, while distress is Type III (aka, not fun at all).

Eustress. The positive stress response, involving optimal levels of stimulation: a type of stress that results from challenging but attainable and enjoyable or worthwhile tasks (e.g., participating in an athletic event, giving a speech). It has a beneficial effect by generating a sense of fulfillment or achievement and facilitating growth, development, mastery, and high levels of performance [APA 2023].

Distress. The negative stress response, often involving negative affect and physiological reactivity: a type of stress that results from being overwhelmed by demands, losses, or perceived threats. Distress triggers physiological changes that can pose serious health risks, especially if combined with maladaptive ways of coping [APA 2023].

Getting started.

Two of the biggest decisions you will make in the PhD are (1) where to go and (2) who to nomintate to your committee. Many personal and academic factors will inform these choices. Think deeply about your unique priorities. I'll propose two points to consider.

Selecting your institution, department, and committee does two things (among others):

1. Buys access to networks. This is the beginning of your social capital in the field. This isn't bought with money, but with your labor as a researcher – and, ideally, with the merit of your ideas. In exchange for doing excellent research for your advisors, you gain access to their networks. Choose your committee with this in mind. Strongly consider which intellectual, professional, and social communities a potential advisor can offer you (and which they cannot). In some cases, you can transcend the limitations of your advisor's networks by forging your own collaborations. I advise against betting on this option, as this is the more challenging route.
2. Buys access to ideas. You are the company you keep. Align yourself with likeminded researchers whom you admire. Good ideas are gatekept; taste and judgement don't grow on trees. Your proximity to talent is extremely important for the quality of ideas you are exposed to. In turn, this will influence the ideas that you yourself develop and pursue in the PhD (and after). Start off on the best foot possible. It is natural (and often positive) to change research directions, explore new communities, etc., but the earlier that you "enter the gates," the faster you reach success.

During this decision-making process, envision your future. Where do graduates from your prospective program tend to wind up in life? Pay close attention to a potential advisor's placements (where their former students work after the PhD). Would you be happy with these placements? If possible, find out how directly involved the advisor was in attaining these placements. For example: Is this professor quick to recommend students to colleagues at prominent universities and companies? Are they generous with letter writing? Are they aware of the state of the job market, including competitiveness and in-demand skills? For some potential advisors, the answers to these questions are no. I advise against working with such people. Relatedly, I strongly encourage those whose work is interdisciplinary to nominate a chair (lead advisor) who is formally trained in and actively publishing in the discipline in which you are most likely to seek employment. Such advisors will be better connected and more culturally aware of the needs of a student in this area.

Technical and intellectual recommendations.

R0. What makes a good scientist? We all want to be good at what we do. In my experience, there are four major traits of a good scientist: (1) technical excellence; (2) creativity; (3) strong written, spoken, and visual communication skills; and (4) good taste and judgement in discerning which problems to work on. In the early stages of your scientific training, you might overindex on technical excellence. Do not forget to invest in these other three traits, especially (4). Devote time to the philosophy of science. In considering my own orientation as a scientist, I found the following very impactful: Birds and Frogs by theoretical physicist and mathematician Freeman Dyson, and the pedagogical reflections of theoretical physicist Richard Feynman (especially his emphasis on explaining concepts in simple terms, and the immortal: "What I cannot create, I do not understand"). As for point (3), perhaps the greatest influence on my writing has been The Elements of Style (Strunk & White).

R1. Every idea you pursue is an opportunity cost. For the most part, you will express your ideas in the form of papers. Papers are your main currency as a PhD student. They are the primary evidence of your capabilities (along with accompanying artifacts, like open-source models, data, and code). They are the primary means by which you can develop skills on a very deep level. They are also a calling card for future employers. They can bring you interview offers and even fame. Never having published in a specific subarea can also get you rejected. Yet, you can only write finitely many papers during the PhD. Among my peers, it is normal to produce 3-7 lead author papers as a PhD student. You can't say yes to every idea (though you may be tempted). Every idea you pursue means leaving other ideas behind. Practice knowing when to say no. Balance your priorities wisely. Passion for the topic is important, as is social good and other broader impacts. However, so is the marketability of your papers when you enter the job market. If your primary goal is employment in a particular subfield, design your papers specifically with high impact, upskilling, and networking in this subfield in mind. Ask yourself: What do you want to be good at, and known for? Whose attention are you trying to catch?

R2. Get used to failure; fail fast and pivot. Early in my career, a senior professor told me that you have to be very comfortable with failure to succeed as a scientist. Experiments fail, hypotheses don't hold, code is often buggy, peer review can be harsh, paper rejections are frequent, grant funding is extremely competitive, and technical interviews are hard to pass. During the PhD, a successful PI consoled me after a paper rejection by saying “I get papers rejected every day!” This is normal, and those who aren't shaken by it are the ones who succeed. Avoid interpreting micro-failures as macro-failures. Do not take rejection as a commentary on your future potential or your self-worth. It is imperative to believe in the potential of your own ideas. You have to be your own champion. At the same time, know when to quit: not every idea is worth a long-term investment. “Fail fast” was one of the best pieces of advice I ever received from a colleague. Read the room and pivot gracefully.

R3. Learn the history of your field and expose yourself to ideas, broadly. Ignorance of prior art will undermine your research. Don't reinvent the wheel. The field moves fast, but research doesn't expire after 5 years: don't fall for recency illusions, read the classics. Attend talks, conferences, workshops, and industry socials on a regular basis. Integrate ideas from diverse sources and across disciplines. Many of the core problems in AI/ML are adjacent to problems that have been studied in mathematics, philosophy, cognitive science, neuroscience, and economics. You never know where you will find inspiration, and it might be very surprising. Thinking is like exercise. Think for the sake of thinking. The more you know the bigger you can dream.

Career development recommendations.

R4. Network and build community from year one. You will learn more, enjoy more, and go further. Network on your own time; don't wait for introductions from your professors. Network vertically and horizontally: your classmates will one day be your internal referrals, your whisper network, your hiring managers, and your greatest source of encouragement. When you hit the job market, you will often be one of hundreds to thousands of applicants per position; even a strong application can get lost in the noise. Unsurprisingly, most people in my network found their first post-PhD job through a first-, second-, or third-degree personal connection (a coauthor, prior internship advisor, professor, someone they met at a conference, a classmate, etc.). But don't be mercenary about it: nurture genuine connections, be giving in return, and focus on learning and growing.

R5. Intern in industry as often as possible. This is the most likely way you will get your first job out of the PhD. This is also an excellent way to build your network, gain practical work experience, get exposure to new ideas, and take the temperature of what research problems actually matter in industry. Do not work in one singular lab for too many years. You need to challenge your own quirks and the quirks you were taught by your mentors. This kind of self-awareness and growth only comes with experiencing, and learning to succeed in, many diverse environments. Having prominent internships on your CV can also significantly boost your legitimacy before you hit the job market. See R4 for one route to finding internships. I personally had three industry internships and one term visiting another university; one internship resulted in an internal referral and one a direct return offer. Many of my peers received full-time return offers from their internships.

R6. Be confident on your own. Community is important, but so is your own self-assurance. Science often feels like a team sport, but the PhD is in many ways a solo sport. You will feel alone a lot. Your fate is ultimately your responsibility and no one else's, even your advisors'. This isn't immediately obvious when you join a PhD program, particularly for those coming straight from undergrad. Institutional and community support will be abundant in some cases, but many of us will not have our hands held. You might find this jarring, feel bitter about it, or struggle to accept the truth. However, it's the beginning of your future: the PhD is meant to produce independent researchers. There will come a point where no one is handing you research projects or employment opportunities, or checking in on your wellbeing. Buffer against the independence (eustress) and loneliness (distress) by building a community of likeminded colleagues. See R2 and R4.

R7. Cultivate a public web presence. We live in a commodified world. Branding is important. Build an attractive personal website. Build a website for every publication, where you can house the paper, code, and all other artifacts, along with a lay summary and attractive visuals. Post about it on social media or write a blog post. Post the videos and slides of your best talks. If you are not averse to social media, build a social media following earlier rather than later. This can help you land job offers, expand the reach of your publications, find job openings, access the RSVP forms for invite-only industry parties, and hear industry news ASAP. Keep your LinkedIn up to date, as recruiters use this platform heavily. When it is time to graduate, the best case scenario is that recruiters and hiring managers reach out to you to chat, rather than vice versa. This is unlikely to happen without an active web presence.

R8. Start studying for technical interviews before you even hit the job market. Studying for technicals is time consuming and important. Studying will take you away from your research. Be prepared with strong time management. Study incrementally over time to prevent overwhelm. Future you will thank you. Failing a technical is a normal experience, even for highly qualified candidates. Be patient with yourself if you do not immediately interview well. There is a wealth of free (and paid) content online to help candidates prepare for technicals. Read the comments left on Glassdoor and online forums about the interview process at the company you want to work for; commenters will often mention the kinds of questions asked. Read subreddits and Medium articles, watch YouTube videos, work on real problem sets, and LeetCode.

R9. Get good at giving talks. Have attractive slides. Learn to speak persuasively and charismatically. This is one of the most important ways to sell your work and yourself as a scientist. Good research + a bad talk is self-sabotage. Learn by observing more senior researchers. Attend as many job talks and conference orals as possible. Start small with practice talks: cut your teeth in local, friendly environments so that you are confident by the time you get your first major oral.

R10. Prepare for an uncertain and rapidly changing world. If you are career-oriented, and especially if you are not born into intergenerational wealth, make hireability a top priority from Day 1. Don't count on the fame of your school to get you a good job; it's not enough. Don't count on your major either; CS/AI is marketable, but it is still not enough. The specific research area you choose (and your impact within it) play a profound role in your future job stability. Your technical skills and the strength of your social capital also matter. Always be thinking ahead to your next step. Speak with those senior to you. Be strategic. Take a bet on what the market needs, and pursue those skills (see R5). Anticipate what will make you useful in this rapidly evolving economy. Be willing to pivot (see R2).

More coming soon.

Personal life, health, and balance.

R11. Invest in your physical health, within your capabilities. Thinking deeply burns calories. Fuel yourself well: one of the most important things you can do during the PhD is to eat as well as possible. Studying, writing, and computing are usually sedantary activities. It is essential to move your body, stand, and stretch as much as possible to offset this. One of the greatest gifts you can give yourself is to invest in your body during the PhD. I took up weightlifting and yoga halfway through the PhD and it was truly transformative. The health of your body supports the health of your mind, and your mind is your strongest asset as a researcher. The psychological, emotional, spiritual, and social benefits of physical activity are also a great way to mitigate distress and carry you through eustress.

More coming soon.