How to leave academia (for science PhDs) • Hacker News Board

Re: How to leave academia (for science PhDs)

#41

Post by jakobe » Thu, Apr 05, 2012, 4:40 PM UTC

kmm wrote: post #32

kentonwhite wrote: post #20

I received my PhD (Physics) in 1999 and went directly to industry. At that time I saw industry as a well paid post doc. My intention was to work a few years and then seek an academic position. In 2001 I had an offer from a top engineering school in Canada to join their faculty. And I said no. Academia attracted me with the promise of intellectual freedom -- the ability to work on what ever problems I wanted to. This…

I'm doing a Masters in theoretical physics and I've got to ask, is it common to first go work a few years before going into academia? It surprises me how little we are informed of these things. Being locked in is my biggest fear for going for a PhD. I don't want to spend the rest of my life on a single topic, perhaps I don't even want to stay in theoretical physics.

I did a masters equivalent degree in theoretical physics, now I'm doing a PhD in computer graphics. A coworker who did his PhD in theoretical physics is now doing a post doc in mechatronic. Many people might specialize more and more, but if you don't like that, no one stops you from changing your field.

Re: How to leave academia (for science PhDs)

#42

Post by bearmf » Thu, Apr 05, 2012, 4:47 PM UTC

eshvk wrote: post #36

bearmf wrote: post #23

Seriously? Learn web development? No one hires PhDs for web development skills. It is only useful if you want to be a pure software developer, not a quant or a data scientist. It might even be a negative when applying for some jobs.

Well they might not be hired to do just web development but I think learning about the web stack is incredibly useful because it helps you get your work out to people. This is important especially in a (early stage) start up where you might have to rapidly prototype solutions and you can't have your own pet designer to whip up visualizations for you. In fact, if you read the blogs of some of the data scientists at tw…

You do not need to be a web developer to write a blog. R, MatPlotLib and Matlab make pretty pictures.

Re: How to leave academia (for science PhDs)

#43

Post by pgbovine » Thu, Apr 05, 2012, 4:56 PM UTC

I know this is cliched, but of the ~30 students who started the Computer Science Ph.D. program at Stanford in 2006 with me, maybe 10 have dropped out. A few of those ten have gone onto successful startups and are now far, far, far wealthier than any of us who remained in the program could hope to be :)

Re: How to leave academia (for science PhDs)

#44

Post by hogu » Thu, Apr 05, 2012, 5:06 PM UTC

telemachos wrote: post #19

> Also, if your degree is in English, the best I can do is point you here (link to Starbucks job page). I know it's supposed to be funny (i.e., it's "just a joke") and I grant I may be sensitive since I have a BA in Comparative Literature (a graduating class of 1 in my major) and a PhD in Classics, but this came across as completely obnoxious and unnecessary. Don't know anything about Humanities? Fine. Don't talk abo…

And also wrong. My wife left academia from philosophy and is pulling in north of 100k right now doing project management stuff. She got there within 3 years, though we're in NYC so it's easier to get more cash here. It was a hard to transition, most employers don't know what to do with a humanities degree, but she was able to do it. A programmer coworker of mine also worked in philosophy, and transitioned to programming and is doing rather well for himself. He did this over the past few years, so long after computer science was well established. Transitioning out of humanities academia into non-academic jobs is difficult, much harder than for those in the sciences, but it's quite doable

Re: How to leave academia (for science PhDs)

#45

Post by hagy » Thu, Apr 05, 2012, 5:10 PM UTC

What about negative anecdotes? Has anyone here regretted his or her decision to leave academia? Particularly, have you found the work less rewarding or interesting?

I ask as a grad student with a background in chemical engineering. I’ve known two professors who’ve done the inverse – industry to academia – and they’ve provided convincing stories for academia. In both cases they found academic research (in engineering and the physical sciences) to be more interesting and personally rewarding than industry.

Is the situation different in software and related industries?

I ask, as I’ve been considering transitioning to such a career myself after grad school (my research is in computational modeling and simulation). My biggest fear is that I’ll find myself in a job that I find boring and will regret my decision. Doubly so, as such a career transition will make it very difficult to reenter my academic field.

Re: How to leave academia (for science PhDs)

#46

Post by dkokelley » Thu, Apr 05, 2012, 5:14 PM UTC

robertskmiles wrote: post #25

> 100 passengers have queued up to board a plane, and are lined up in the order of the seats on the plane (n=1..100). However, the first person lost his ticket and selects a random seat. The remaining passengers will occupy their assigned seat if it is available, or a random seat otherwise. What is the probability that passenger 100 sits in seat 100? It really worried me that I couldn't figure out how to work that ou…

I have no idea if that is correct, but my initial intuition was that it would be a high percentage like you found. However, now that I'm thinking about it, I believe the answer is closer to 50%.

When passenger 100 begins to board, there are only two possible configurations he can find the seats in: seat 1 available or seat 100 available. If Passenger 1 sits in seat 100, then passengers 2-99 will sit in their correct seats, leaving only seat 1. If passenger 1 sits in seat 1, passengers 2-99 will sit correctly and leave only seat 100 for passenger 100. If passenger 1 sits anywhere else (say, seat 50), passengers 2 - 49 will fill correctly, passenger 50 will have to sit in seat 1 or 51-100. The moment a passenger randomly selects seat 1 or 100, the remaining passengers will sit correctly (except for passenger 100). It is impossible for any seats other than 1 or 100 to be left for passenger 100 to select.

From looking at it, I don't believe there is any bias for the end configuration to leave seat 100 open more than seat 1, and I believe that in an interview situation, you would be expected to intuit this by "talking it out". If I'm right (and I could still be very wrong), this is more of a probabilistic brainteaser than a programming interview question.

Re: How to leave academia (for science PhDs)

#47

Post by Rickasaurus » Thu, Apr 05, 2012, 5:16 PM UTC

bearmf wrote: post #40

Rickasaurus wrote: post #37

Just tossing in that Finance isn't your only option for a healthy PhD paycheck in NYC. With exactly those same skills you could work in silicon alley or at a company like mine which does analysis on bank (or other corporate) data. Also, you don't have to sell your soul to work outside of Academia. For example, we find terrorists and drug lord hiding cash in USA banks (Anti-Money Laundering) using machine learning. It…

Where are all this high paying jobs? Average data scientist salary at glassdoor is probably less than 100k.

Glassdoor doesn't seem to give me any salary results for "data scientist" in NYC.

Re: How to leave academia (for science PhDs)

#48

Post by bearmf » Thu, Apr 05, 2012, 5:26 PM UTC

Rickasaurus wrote: post #47

bearmf wrote: post #40

Earlier quoted context omitted.

Where are all this high paying jobs? Average data scientist salary at glassdoor is probably less than 100k.

Glassdoor doesn't seem to give me any salary results for "data scientist" in NYC.

Yeah, sorry, these hits were some other scientists. So what is the pay like in NYC?

Re: How to leave academia (for science PhDs)

#49

Post by dxbydt » Thu, Apr 05, 2012, 5:28 PM UTC

robertskmiles wrote: post #25

> 100 passengers have queued up to board a plane, and are lined up in the order of the seats on the plane (n=1..100). However, the first person lost his ticket and selects a random seat. The remaining passengers will occupy their assigned seat if it is available, or a random seat otherwise. What is the probability that passenger 100 sits in seat 100? It really worried me that I couldn't figure out how to work that ou…

Its 50%.

For 3 passengers the seating possibilities are {123,213,231,321}, so 3 gets to sit on seat 3 50% of the time.

For 4 passengers the seating possibilities are {1234,2134,2314,2341,2431,3241,3214,4231}, so 4 gets to sit on seat 4 50% of the time. From there you can do a proof by induction.

If you like code: ----

     import util.{Random=>rng}
     import collection.mutable.ArrayBuffer

     object passengers {
       def main(args:Array[String]) = {
         val n = args(0).toInt
         val simulations = args(1).toInt
    
         val happyLast = (1 to simulations).map(_=> {
      
           val seats = ArrayBuffer.fill[Int](n+1)(0)
           val taken = new ArrayBuffer[Int]
           val all = (1 to n).toSeq
           val seat  = rng.shuffle(all).head
           seats(seat)  = 1
           taken    += seat
           (2 to n).foreach( p=> {
              val seat = seats(p) match {
                   case 0=> p
                   case x=> rng.shuffle(all.diff(taken)).head
              }
              seats(seat) = p
              taken    += seat
           })
           seats(n) == n
         })
         println( happyLast.filter(x=>x).size*1.0d/simulations )
       }
     }

----- >scala -cp classes passengers 100 100000 >0.4997

Re: How to leave academia (for science PhDs)

#50

Post by waterhouse » Thu, Apr 05, 2012, 5:34 PM UTC

Jeeeeesus christ, that grep is ugly.

  grep "[0-9]\{3\}[-]\?[0-9]\{3\}[-]\?[0-9]\{4\}" filename.txt

The default behavior of grep appears to be to not treat {}?() as special characters. Which is generally boneheaded in my experience, and certainly in this case. An irritated man will use egrep--short for grep -E, "extended regexes"--i.e. sane regexes:

  egrep "[0-9]{3}-?[0-9]{3}-?[0-9]{4}" filename.txt

This is just a direct translation. If I wrote the regex myself, I would probably also recognize spaces as delimiters (as in "999 999 9999"), and maybe note that the area code might be missing or have parens around it, etc.--depending on just what I was trying to parse. I dunno, maybe he's parsing regular machine output and his original regex would suffice, so I won't go farther than an equivalent translation. But I will, for kicks, mention this:

  egrep "([0-9]{3}-?){2}[0-9]{4}" filename.txt

(Note that - is not a special character in either grep or egrep, although command-line programs in general treat specially any argument beginning with -, so it is a good idea to use [-] instead of - if that happens to be the first character of your string. But after that, no.)

Also, for general purposes, grep -P (Perl regexes) is best of all, because it can do *? non-greedy matching. A really irritated man has aliased pgrep to "grep -P --color=auto" in his bash startup file, and has also (perhaps dangerously) aliased grep to it as well. (That breaks fewer things than altering GREP_OPTIONS.)