Blog

CS industry internship during PhD

Aug 14, 2021

Last summer, I was a research intern at a tech giant company. When new interns asked me about what to expect, here's what I frequently said.

How the internship differed from the PhD

Evaluation criteria. In school, the newer and fancier the technique, the better. In the company, the opposite was true: the less risky and simpler, the better, as it would cost less. In school, being successful involves impressing other scientists, who judge us by how creative we are. In the company, success requires impressing customers, who may not be interested in buying our new products. So, superstars in one place might not instantly perform equally well in the other place.

Incentives for planning. In school, the more ambitious our proposal looks, the better. So we tend to propose more than what we can realistically achieve. In the company, the proposal needs to be more conservative. The more reliably we deliver the promised milestones, the better. My internship mentor even suggested proposing a plan so simple that I would feel confident completing it in half of the proposed time. Doing this would make both me and my manager happy.

Forced work-life balance. In the company, no one was reachable after hours and on weekends. No messages received, no meetings scheduled. Even if I wanted to work after hours, it was hard to make progress because I couldn't find anyone to answer my questions. In school, however, many meetings are scheduled in the evenings or on weekends. Many academics are proud of working after hours since they are supposed to love their research and devote all their time to work. If we take breaks, our research may slow down and lose to competitors who work longer. In contrast, taking breaks in the company was fine as long as the project delivered on time.

Financial aspects. I was paid much better in the company than in school. For the first time in my Ph.D., it crossed my mind that I didn't want to waste my energy deciding whether a food delivery was worth the money.

How the internship felt similar to the PhD

Quick prototyping. In both school and the company, I was advised to build prototypes quickly before devoting significant effort to carrying out a plan. The only main difference was in what the prototypes were supposed to convey. In the company, the prototype was for convincing other people that my idea could turn into a product. In school, on the other hand, the prototype is used to convince myself. So, it's more about testing hypotheses and probing whether the hardest subproblems are tractable or not.

Talking to people. In school, we talk to people for many benefits: practicing communication, building connections, getting inspired, seeking feedback, advertising our work, keeping up with the state of the field, etc. In the company, talking to people has similar benefits. When looking for a project, I chatted with people in different positions to understand the company's value system. When preparing the final presentation, I gave practice talks to others who provided valuable feedback that I could not have imagined on my own.