Ok, so this is for a very focused crowd, I guess, but I know a lot of people who will be in my shoes, and I feel like I learned something, so I want to write about it.

I guess first we have to do the “industry or academia???” question. This one is simple. If your answer is “Academia, hell yes, 100% for sure”, do academia; otherwise, do industry.

Ok, so you’re doing industry. Here’s the first question you might want to ask yourself: what kind of industry job? Here are some options:

Research Scientist

This is an increasingly rare job, showing up only at some of the megacorps like Microsoft Research, FXPAL, and somehow Yahoo. Do this if you actually wish you were in academia. At MSR, it’ll be close to academia; at other companies, it’ll be farther, but you can still make up new things. I interviewed for one of these; it involved me giving a talk about my research and then talking with 4 research scientists/engineers on the team one by one. At the end, I realized I didn’t know enough machine learning to work with this team (the Machine Learning and Vision team).

User Experience Researcher

These people do research with people all the time. Interviews and surveys are very common, but you might also do focus groups, workshops, intercepts, usability studies, and some stats. You also have to have some skills in communicating, and you probably have to not mind convincing people that your work is real. I interviewed for a few of these, and they usually involved me giving a talk about some research I’ve done, then I talked with 5-7 people one by one. At one of them, they gave me a short problem, something like “You work at (hypothetical company X), and they have this new broad question. How would you narrow it down into a research plan?” and I had an hour to think about it and then give a 20-min presentation. I don’t really know how these interviews are evaluated.

Software Engineer

You like writing code in a team. You were frustrated at all the terrible code you wrote throughout grad school. You like organizing things. If you think you want to be a SWE, you probably shouldn’t go to grad school, but if you graduate and find that that’s what you want to do, then by all means do it! You will always have a well-paying job and be treated like a minor royalty. Especially if you’re good.

UI or UX Designer

If you know you like to do this kind of thing, go for this job! I don’t know anything about it, really.

Project Manager

Do you like to have 500 tabs open and communicate with everyone all the time? And, like, deal with stuff blowing up and figuring out how to coordinate a bunch of screaming cats to get their stuff together to actually get a product out the door? Do you think grad school was too slow and quiet? PMing is for you! God bless you. I could never handle this job.

Data Scientist

This… can mean a lot of things. But most of them are in demand! It can mean:

• Product analytics person - answer questions about your product
• Quantitative UX person - answer questions about your users
• Machine learning engineer - building “data products” (where “data products” means roughly anything where you are doing any ML)
• The first person at your startup who works with data for more than one-off things
• Some combination of the above
• Probably something else too

These were most of my interviews. They usually started with 1-3 phone screens, then sometimes a “homework assignment”, then a daylong interview with 5-6 people, which were some combination of:

• Programming
• SQL
• Experimental design
• Big-picture thinking (~ “we want to do X with our website, how would we do it?")
• Machine learning or modeling
• Collaborating with other people/teams
• Lunch