Academia and industry aren’t that different

Let’s break down some of the parallels between academia and industry by stages of job hunting & applications, interviewing, and actually doing the job.

Note: My perspective is biased based on coming from a research background and going into a full-time research role as an Applied Scientist.

---

During the job hunting & application process

Finding grad advisors vs. managers

The greatest way that you increased your chances of acceptance to a grad program was if grad advisors were taking on grad students that application cycle.

For PhD programs, principal investigators (PIs) may extend offers to a certain number of grad applicants a year. Or may not, if there is no need or if their department did not provide them grad student “seats” that year.

In industry, this is called “headcount”. This is the number of new full-time roles that a hiring manager is given per fiscal quarter or year. Note that there can also be different types of headcount in companies.¹

Figure out who has open or upcoming headcount

Just like when applying to grad school, you can figure out who has open or upcoming headcount by:

• talking with grad students or the lab manager in the lab = networking with people in the company and/or specific team
• talking to other advisors who could become your collaborators = networking with hiring managers who may know of other hiring managers who are actively hiring

Also, it’s important to keep in mind that industry follows different timelines. For many STEM PhD programs, the application cycle is Sept-Dec. For industry, figure out when hiring managers will hear about their headcount allocations, when there are naturally more open headcount due to budgeting, and when hiring slows (e.g., winter holiday season).

Identifying lab focus vs. team focus

Back in 2018: When I was considering which labs & PIs to apply to, I realized that it was hard to figure out what a potential advisor was actively working on and interested in pursuing at that moment. When you look at a lab’s list of publications, many of those papers may have been recently published but the project itself might have ended years ago. I realized that by the time that I applied, it was possible that the lab may have shifted in research focus long after a paper was published or its website was last updated.

Rather than only reading papers, I also looked up the PI on YouTube to watch their recent research talks, checked if they’d written any medium or substack posts, and also checked their science Twitter/X. (Today, I would recommend Bluesky.)

I also asked the PI what they were interested in, if we were able to talk in advance of me applying. But it depends on the PI and is not always possible for applicants. (If they say that they do not talk to applicants in advance of applications, don’t ask them for a chat.)

But this is possible for hiring managers. If you’re able to find them online, ask them to a virtual coffeechat. How do you find them and also figure out what research area(s) that they are hiring for? Again:

• read medium/substack/personal websites
• watch YouTube research talks
• LinkedIn - hiring managers very often post about open full-time roles and most will often post if they are recruiting for internships
• read relevant papers - many industry researchers are still publishing!

---

During the job interview process

The interview starts in the first interaction

When I saw PhD candidates begin to interview for postdoc positions or postdocs begin to interview for faculty positions, it was clear that their audience was also basing their decisions on interactions that had been building long before then—meeting at conferences, presenting at invited lab meetings, etc.

In fact, advice that I received early on in grad school from The Grad Advisor was that “Every talk after the 3rd year is the job talk”.

In industry,

The interview starts in your first networking conversation with anyone on your target team.

It’s high pressure², but it makes sense. If this conversation does not go well, they are likely not to introduce you to their hiring manager or (even if you do officially interview with the team) they will likely not endorse you in the review process.

Treat every interaction genuinely and professionally.

When to ask which questions

It’s important to prioritize your work-life balance. But you don’t need to prioritize asking about this in the very first interview that you have with a PI or hiring manager. Just like in academia, the focus on your first interviews should be able your research.

After you’ve received an offer, then you can request additional chats. Sometimes the informal, follow-up chats are the most insightful about lab culture and PI mentorship. But again, you want to save asking these questions until after you’ve secured an offer.

Same goes for industry. Wait until you’ve received an offer to ask about logistics like PTO, flexible hybrid work, benefits, etc. I’ve also requested office visits before to get a sense of my potential work environment. Unlike in academia, industry interviews move fast and they expect you to make decisions quickly. So be prepared to move quickly once you have been made an offer.

Competence is different from growth potential

When I was a grad student, I watched many postdocs give job talks for faculty positions. (Take the following observation with a healthy amount of salt, because I did not pursue this route for personal reasons³ and therefore, do not have faculty perspective.) The reason that I perceived that a candidate did not receive an offer was that even if they were more than competent as a postdoc, they did not show growth potential for building a longterm research program in the department.

I’ve seen this pattern hold for academics interviewing for industry roles. Even if you are clearly brilliant with an impressive research background, it’s imperative to show that you can:

• flexibly think about new research problems that come with the job (e.g., how GenAI models may develop in the years to come)
• build a research program (expertise & skillset) that addresses longterm problems and generates longterm solutions—in other words, show that you have growth potential beyond an internship or contractor role
• learn new skills or domain knowledge that may be outside your current research

Most importantly, it’s a deal-breaker if you’re unable to show that you’d be able to get along with everyone on your direct team, collaborating with cross-functional teams, leadership, and client stakeholders. In academia, faculty are selecting who they want as a hallway neighbor for upwards of 4 decades. The same goes your teams will be having weekly or daily meetings with you for n years.

---

During the job

Minimally publishable vs. minimally actionable

How do you know when a paper is done?

In academia, “minimally publishable” is the amount of original research that can standalone as a unit of meaningful contribution as e.g., a conference paper, peer-reviewed journal publication, brief report. In grad school, this personally helped me conceptualize how many studies in a paper were needed to meaningfully address a research question when, of course, the original research question led to more questions, which led to more questions, etc. In other words, while there was always more to the story, the concept of “minimally publishable” helped me constrain what to include as the beginning, middle, and end of a paper — i.e., the research story to be shared out to the broader scientific community.

How do you know when a project is ready to be shared out with your manager, team, to stakeholders, or more widely within the company? When a meaningful decision or action can be made.

The main constructive feedback that I received from my manager in my first year was to share out findings earlier. I’m a perfectionist, and this 100% works against me in my current job. Sharing out earlier does not mean messier. It means learning how to identify when I have enough insights that decisions can be made by another team or leadership. Learning what is minimally actionable involved integrated old & new skills:

• Piloting — Running a series of incremental studies (i.e., our standard pilot studies!) in which results can regularly be shared out along the way
• Strategic sense of time — Results are more meaningful where there are decisions to be made and become meaningless when that time has passed. Developing a strategic sense of time involved learning (1) critical project timelines, related learning general business knowledge; and (2) how to read people and their sense of urgency, which is more social than you’d expect
  • A personal heuristic is that if 1 important person says the same thing twice, I’ll move on it. If 2 different people say the same time, I’ll move on it. Both signal genuine business need and the start of team/company momentum towards real business impact.
• Flexible communication formats — Some decisions are more important than others. This will change the pace in which you share out results. Sometimes, a polished slide deck will be worth the time and effort, and other times a minimally written slack canvas will have the same business impact.
  • Note: Vibe-coded slides look amazing, but don’t support collaboration in the form of “comments” and co-editing/writing in the same way that slack canvases, google docs, and word docs do.

Research gaps vs. market trends & data gaps

Literature reviews were critical throughout all our time in grad school to (1) understand the status of the field; and (2) identify research gaps. It was the first place that we started for most research projects. It can feel a little or a lot mundane, but personally I think is one of the most hard-won skills from the PhD. Why?

In the first year (or more) of your industry job, it’ll likely be critical for you to be good—and later, great—at the things that you were hired to do and to know. This is pure execution. But eventually, your real value will be in what you don’t yet know and don’t yet know how to do.

Your real value will be in identifying which aspects of what you don’t yet know is also what the market doesn’t yet know.

Eventually the market you were originally hired into will shift. Knowledge & skills will become outdated. You can catch up. But it’s more impactful to you and your team/company if you can identify the direction that the market is shifting before you all need to catch up. You’ll hear that executive leadership will also want your company to be leading/defining/“winning” the market. Your value comes from sensing the wind shifting, identifying the direction, and bringing your team the data needed to be the first to lead in that new market space.

While all forms of jobs are valuable, I personally think that PhD students excel in flexibly shifting from execution to strategizing. Specifically, in longterm end-to-end research cycles. This directly builds on the idea of “Competence is different from growth potential”.

Here’s an except from Phase 4: Learn the market trends & data gaps from The first 365 days of your first job in industry:

One of the most important things that I realized is that everyone is learning about AI at the exact same time that I am. (My work involves evaluating GenAI models.) This means that there’s no standard for expertise in what one should know. In fact, I learned that my true value in my job is identifying what we don’t yet know in the market and figuring out ways to “close the gap”.

Does that sound familiar?

Academia is all about identifying research gaps! It’s the bread and butter of our literature reviews. Leverage this hard-won skillset that you developed in the last 5+ years of your academic training.

Communication is critical

There are two critical axes to consider: (1) horizontal = cross-functional audiences; and (2) vertical = executive level. This kind of audience design is very common to academics, who present to many different scientific communities at conferences (horizontal axis). We also have a lot of experience communicating results to undergrads, masters students, PhD peers, and professors (vertical axis)

However, I found that I had significantly less experience with the executive level and had to learn this quickly.

Based on these axes, consider how you can adapt your communication based on your audience’s

• goals - you might explore new user personas, how to save or make more resources (budget, GPUs, time), close on a deal, common ground a new project between teams
• communication style - try out wire framing, vibe coded prototypes, PRD, slides
• stage of product/model lifecycle - think about training dev, iterations, go/no-go

For example, I often find that the higher in leadership my audience is, I adapt my slides to be shorter, higher-level, and with even more clear takeaways related to results, action items, and (if relevant) estimation of resource costs. I also tend to use slides for image examples which is relevant to my current role evaluating GenAI images or schematics but also because they are faster to comprehend. I also often copy-paste my schematics from my slides into a short slack canvas later on. This also makes your slides more memorable when building a research narrative or pitch over the course of a few weeks or months of talking to leadership.

Or, for example, with engineering and design teams, I often mock up designs in a scrappy visual format like keynote or vibe-coded prototypes. Sometimes I do this in advance or even on the spot with teams. This format makes it easier for engineering and design teams to create the real, productionized version of my mockups. I also often focus on detailed user workflows and potential user personas (e.g., demographics like how users fit into various customer segmentations). A breakdown of the requests into order of priority is also critical to cross-functional collaboration.

---

AI was not used to write, edit, or brainstorm this post. All mistakes and em dashes are my own.