Interviewing Ph.D.’s

What I look for when hiring a Ph.D.

Robert McKeon Aloe
3 min readAug 25, 2020

I have interviewed mostly Ph.D.’s for computer vision / data science roles this past year. I have a great committee that quizzes them on all sorts of things from problem solving to coding to cultural fit. I talk with them in a general way and eventually, I settle on my two questions that help clarify the strength of the candidate:

  1. Tell me about your dissertation.
  2. What was the piece of your dissertation you regretted the most?

Tell me about your dissertation

A dissertation takes years to finish. It is a summation of the cheapest work you’ll ever get paid for because the hours to pay ratio is huge. There was such a motivation to finish my dissertation sooner so I could get a job to pay me six times my stipend. This caused a lot of hard work.

As such, my dissertation is my baby. I cherish it; I cherish the time. I cherish the experience, and I can look back fondly on the pain and joy. As a result of explaining the topic to my advisor, my parents, my wife, my peers, at conferences, and in journals, I’ve gotten pretty good at convening the concepts in simple terms. I’ve also gotten good at explaining the difficult parts.

Dissertation Regret

What piece of your dissertation did you regret the most when you finished? If the answer is nothing, then you haven’t dissertated enough. In my experience, there were many things I would have done differently if I could do them again. I would have collected less data but better data. I would have taken more time to do failure analysis. I would have built the next prototype.

However, if I would have done those things, I’m positive that I would not have finished on time, I would have found other things to regret, and I would not have failed enough. This failure was required because the learning of a Ph.D. is how to do the work of research.

Dissertations are about failing and learning.

Research as a whole is dominated by failure. The months of failure, bugs, bad DOE’s, compliance issues, and delays completely outweigh the few weeks of reprieve when things seem to go right. All my research has ever been is grinding work until the last day or two when the final result comes out, and I compile it into a presentation or production code.

Then, having solved some grand problem, I move on to the next.

So, tell me about those experiences. Those learning experiences are more valuable than the successes because your success was built on your failure. Then I can see how you respond to the inevitable failure you will encounter working with me. Past success doesn’t guarantee future success; humility learned through trial by fire will help mature your professional development.

--

--

Robert McKeon Aloe

I’m in love with my Wife, my Kids, Espresso, Data Science, tomatoes, cooking, engineering, talking, family, Paris, and Italy, not necessarily in that order.