Interviewing

Sometime ago, the PhD students at Rutgers organized a talk and asked me to speak about job interviews. Most knew about the academic job market, and really wanted to know about jobs at (research) labs and companies. Jobs in labs fall into at least 3 categories:

• one writes a research paper and then shops around to see if someone in the company will find it useful.
  
• one first builds something useful and later writes the paper.
  
• one just builds.

The companies differ on what building means:

• code enough to do the plots in the paper.
  
• code enough to prototype, GUI included, and shop around.
  
• code the prototype and get others to use it internally (in the company, community).
• code to release for (paying?) customer use.
  

Seen from the interviewer point of view, one can tell when someone has done their homework. When asked what would a candidate do at the job, one of the common responses is, "I want to be close to the data". I don't usually know what that means, and as Graham Cormode (paraphrased) quips, "When one says they want to be close to the ocean, do they want to watch the waves or surf?".