- #1
gpsimms
- 30
- 1
Hi there,
TL/DR - I have a community college job with a sufficient salary. Should I be happy with that + freetime or should I try to get a job which really uses my PhD? If I go with happy freetime option, are any of the projects I propose interesting AND feasible? If I am too lazy to follow through on any projects, is it a bad thing?
-----------------------------------------------------------------------------------
I didn't read through previously posted threads because it seems like this kind of advice is pretty personal and case-by-case. Regardless, I apologize for not reading through threads and if someone would rather point me to a good one, please do.
Anyway: I am a few months from wrapping up a PhD in mechanical/aerospace engineering. My concentration was combustion chemistry, particularly of real aircraft fuels. I am an experimentalist, and I setup flow-tube reactor studies of combustion products. I performed measurements with gas chromatography (GC) and molecular beam mass spectrometry (MBMS).
I am not passionate about combustion, and will not be heartbroken to leave the field entirely. I do love science, and have become quite handy as an experimentalist, and wouldn't mind keeping active in experimental research generally (whether it is combustion or something else). I guess it seems like I have picked up skills where I could be an analytical chemist-type person.
I have gotten a job. I was offered a full-time position at the local community college. This is a permanent position, and I can make about $60k in 9 months, with the option to add a few more summer classes, etc. and possibly get an extra ~$10k or so with overload and/or summer courses.
The money is only important in the sense that I will make enough at this job (factoring in my wife's salary, local cost of living, etc.) that I could work this job indefinitely and we would be able to pay off our loans and live relatively comfortably.
I do like teaching, and always have. However, I can easily imagine the possibility that community college teaching could get old. It does have some perks, though:
*Like I said, salary is sufficient, and we do like the area.
*My degree is aerospace engineering, with an MS in applied math. As such, the community college has basically told me that while I am officially to be an associate professor in math, in the future I will likely have an opportunity to teach engineering, computing, chemisty, or physics as well. Teaching 5 math classes where 3 are pre-calculus sounds kind of crappy. But, teaching an interesting math class, an interesting chemistry class, an interesting physics class, a coding class, and running a lab class sounds pretty nice, for example.
*This is local to the university from which I am graduating, and it is possible I may be able to keep a working relationship with my advisor so that I can keep half a toe in research if it still interests me. (Additionally, because I will be "active" still in research, I don't think I will be closing the door on applying for research jobs in the future).
*I have taught at the high school, community college, and university level (as a TA), and I am pretty sure that the community college schedule will leave me underworked. That means, particularly if I take summers off, that I will have some free time. I have a daughter, and will probably have another kid or 2, so free time to be a good dad is worth a lot, too.
-------------------------------------------------------------------------------------
OK, if anyone wants to give advice on whether private industry, or government labs, or postdocs, or whatever is the way I should go, feel free. It is also possible that I could, down the road, get a professorship at a local liberal arts college, where I would not have a heavy expectation of producing research. Does anyone have any thoughts on that lifestyle?
-----------------------------------------------------------------------------------
Another point was to talk about a few random interests, and see if people thought they are interesting avenues to go down...
1.) A voice-to-text app combined with machine learning (if necessary?) to write doctors' notes for them
So my wife is a physician, and by far the worst part of her life is the documentation. It seems to me like the conversations usually follow a pattern (for each particular sub-specialty). Isn't machine learning and voice-to-text well suited to solve this problem? Now, I am *not* a computer scientist. I use computation to run combustion models and to process data. I know a little bit about image processing, nothing about audio processing. I know nothing about machine learning. I know nothing about apps or how to code them. Forgetting about privacy and permission issues for a moment, is this the kind of problem someone with my skills could learn enough about to solve? I am imagining having fake conversations with my wife, looking at the notes she would write, and teaching a computer how to pull the important parts out of a conversation to make the note.
2.) Image processing and/or sensor technology to assist officials
I took a couple of week mini-course on image processing. My skills are extremely rudimentary. But. Like most americans I love sports and I hate the refs. The sport that has most successfully implemented automation of officials is definitely tennis. Bad line calls are corrected immediately with a near 100% success rate. Baseball should be doing this with balls/strikes but of course they are refusing to use the technology right in front of them.
Football, soccer, hockey, etc. are much more complicated, and gains from image processing would be marginal. However, there are cases where it could be useful: determining, using multiple camera angles and precise definitions of the camera locations, the exact location of the ball at a particular moment of time, for example. Determining the moment that a foot touches the ground, or out of bounds, etc.
Eventually, machine learning could probably be taught more complicated calls: a hold vs. a legal block in football, a charge vs. a block in basketball, etc.
Again, as someone with basic coding experience (MATLAB, python, fortran), is this a problem that a month out of the year of free time would be enough for me to solve?
Another approach would be to use sensors/accelerometers embedded in the ball, shoes etc. to get precise locations that way. I have no experience with those types of devices, but have some general experimental engineering skills as I have mentioned above. There must already be someone trying to get location devices inside a football without affecting the way the football feels, etc., right?
3.) Sports analytics/prediction
Obviously, this market is pretty well saturated and I doubt I have any revolutionary idea which will make me a vegas-beating millionaire. But, like I said, I do like sports, and I think I have a semi-novel idea. As far as I can tell, all the prediction machines for basketball just take average efficiency margin and average tempo and pump out a score. But we sportsfans know that matchups matter, and we should be able to predict team Y which is especially reliant on transition baskets will struggle against team X which is especially good at limiting its opponents transition opportunities. Data on those metrics exist, so I would like to think we can make stronger predictions. In fact, synergy sports puts out tons of play-by-play data (e.g. team X takes 20% of its shots in the half court as spot-up 3s, which it makes at a 38% clip, while team Y only allows opponents to take 13% of its shots as spot-up 3s). So I am proposing simulating the play-by-play of the entire game to predict the outcome. Because these data exist, it feels like this is also a problem well-suited for machine learning.
Again, can a non-CS PhD with a few months of free time even think about tackling this type of problem?
4.) Carpentry/homebrewing
I probably will do a bit of this regardless, and this one is not helping humanity or (likely not) making money, but it's become a fun hobby. My house is on an acre of land with a good garage, which we don't use for cars. I have started converting the garage into a pretty cool bar. I have a slight inkling to build another structure and use all my lab skills (plumbing, PID temperature control, etc.) to make it into a brewery. I don't think I'm a good enough brewer to "go pro," and it feels like the brewery market must be near saturation. But it's just a fun way to spend time.
5.) Online poker
I have always liked poker, and am a marginally good player. I have some statistics on limited online play:
in about 120 sittings in online play, I have a 15% return on investment over a 6 month period, but it is only 7.5% over the last 1/3 of that, or so. I think these statistics are not very robust because the sample size is low. Also, these are cash games on the order of $5-$10 buy-ins, so an average day would be sit down with $10, leave with 10.70 after an hour or something.
However, I've never invested real time in working on it, and I've definitely never employed analytics. Anyone out there able to comment on my "beginner" stats, and whether using computing to learn about optimal strategy is a window to significant improvement? EDIT Or are the robots already about to take this space over? Or is it so saturated with real quantitative thinkers and not enough weak players that it will be too competitive to make any gains? 6.) Will I actually do any of that stuff anyway?
This is my main fear of the free time route: all of these ideas are probably harder and less interesting than I think. I doubt I'll get super passionate about any of them. I might just waste all my free time and be a less interesting person than I would have been if I became a legit scientist. OTOH, maybe all that "wasted" time will just be hanging out with my kids and family more, and making them into the best human beings I can.
Thanks for reading all that! I'd love to hear any and all thoughts.
TL/DR - I have a community college job with a sufficient salary. Should I be happy with that + freetime or should I try to get a job which really uses my PhD? If I go with happy freetime option, are any of the projects I propose interesting AND feasible? If I am too lazy to follow through on any projects, is it a bad thing?
-----------------------------------------------------------------------------------
I didn't read through previously posted threads because it seems like this kind of advice is pretty personal and case-by-case. Regardless, I apologize for not reading through threads and if someone would rather point me to a good one, please do.
Anyway: I am a few months from wrapping up a PhD in mechanical/aerospace engineering. My concentration was combustion chemistry, particularly of real aircraft fuels. I am an experimentalist, and I setup flow-tube reactor studies of combustion products. I performed measurements with gas chromatography (GC) and molecular beam mass spectrometry (MBMS).
I am not passionate about combustion, and will not be heartbroken to leave the field entirely. I do love science, and have become quite handy as an experimentalist, and wouldn't mind keeping active in experimental research generally (whether it is combustion or something else). I guess it seems like I have picked up skills where I could be an analytical chemist-type person.
I have gotten a job. I was offered a full-time position at the local community college. This is a permanent position, and I can make about $60k in 9 months, with the option to add a few more summer classes, etc. and possibly get an extra ~$10k or so with overload and/or summer courses.
The money is only important in the sense that I will make enough at this job (factoring in my wife's salary, local cost of living, etc.) that I could work this job indefinitely and we would be able to pay off our loans and live relatively comfortably.
I do like teaching, and always have. However, I can easily imagine the possibility that community college teaching could get old. It does have some perks, though:
*Like I said, salary is sufficient, and we do like the area.
*My degree is aerospace engineering, with an MS in applied math. As such, the community college has basically told me that while I am officially to be an associate professor in math, in the future I will likely have an opportunity to teach engineering, computing, chemisty, or physics as well. Teaching 5 math classes where 3 are pre-calculus sounds kind of crappy. But, teaching an interesting math class, an interesting chemistry class, an interesting physics class, a coding class, and running a lab class sounds pretty nice, for example.
*This is local to the university from which I am graduating, and it is possible I may be able to keep a working relationship with my advisor so that I can keep half a toe in research if it still interests me. (Additionally, because I will be "active" still in research, I don't think I will be closing the door on applying for research jobs in the future).
*I have taught at the high school, community college, and university level (as a TA), and I am pretty sure that the community college schedule will leave me underworked. That means, particularly if I take summers off, that I will have some free time. I have a daughter, and will probably have another kid or 2, so free time to be a good dad is worth a lot, too.
-------------------------------------------------------------------------------------
OK, if anyone wants to give advice on whether private industry, or government labs, or postdocs, or whatever is the way I should go, feel free. It is also possible that I could, down the road, get a professorship at a local liberal arts college, where I would not have a heavy expectation of producing research. Does anyone have any thoughts on that lifestyle?
-----------------------------------------------------------------------------------
Another point was to talk about a few random interests, and see if people thought they are interesting avenues to go down...
1.) A voice-to-text app combined with machine learning (if necessary?) to write doctors' notes for them
So my wife is a physician, and by far the worst part of her life is the documentation. It seems to me like the conversations usually follow a pattern (for each particular sub-specialty). Isn't machine learning and voice-to-text well suited to solve this problem? Now, I am *not* a computer scientist. I use computation to run combustion models and to process data. I know a little bit about image processing, nothing about audio processing. I know nothing about machine learning. I know nothing about apps or how to code them. Forgetting about privacy and permission issues for a moment, is this the kind of problem someone with my skills could learn enough about to solve? I am imagining having fake conversations with my wife, looking at the notes she would write, and teaching a computer how to pull the important parts out of a conversation to make the note.
2.) Image processing and/or sensor technology to assist officials
I took a couple of week mini-course on image processing. My skills are extremely rudimentary. But. Like most americans I love sports and I hate the refs. The sport that has most successfully implemented automation of officials is definitely tennis. Bad line calls are corrected immediately with a near 100% success rate. Baseball should be doing this with balls/strikes but of course they are refusing to use the technology right in front of them.
Football, soccer, hockey, etc. are much more complicated, and gains from image processing would be marginal. However, there are cases where it could be useful: determining, using multiple camera angles and precise definitions of the camera locations, the exact location of the ball at a particular moment of time, for example. Determining the moment that a foot touches the ground, or out of bounds, etc.
Eventually, machine learning could probably be taught more complicated calls: a hold vs. a legal block in football, a charge vs. a block in basketball, etc.
Again, as someone with basic coding experience (MATLAB, python, fortran), is this a problem that a month out of the year of free time would be enough for me to solve?
Another approach would be to use sensors/accelerometers embedded in the ball, shoes etc. to get precise locations that way. I have no experience with those types of devices, but have some general experimental engineering skills as I have mentioned above. There must already be someone trying to get location devices inside a football without affecting the way the football feels, etc., right?
3.) Sports analytics/prediction
Obviously, this market is pretty well saturated and I doubt I have any revolutionary idea which will make me a vegas-beating millionaire. But, like I said, I do like sports, and I think I have a semi-novel idea. As far as I can tell, all the prediction machines for basketball just take average efficiency margin and average tempo and pump out a score. But we sportsfans know that matchups matter, and we should be able to predict team Y which is especially reliant on transition baskets will struggle against team X which is especially good at limiting its opponents transition opportunities. Data on those metrics exist, so I would like to think we can make stronger predictions. In fact, synergy sports puts out tons of play-by-play data (e.g. team X takes 20% of its shots in the half court as spot-up 3s, which it makes at a 38% clip, while team Y only allows opponents to take 13% of its shots as spot-up 3s). So I am proposing simulating the play-by-play of the entire game to predict the outcome. Because these data exist, it feels like this is also a problem well-suited for machine learning.
Again, can a non-CS PhD with a few months of free time even think about tackling this type of problem?
4.) Carpentry/homebrewing
I probably will do a bit of this regardless, and this one is not helping humanity or (likely not) making money, but it's become a fun hobby. My house is on an acre of land with a good garage, which we don't use for cars. I have started converting the garage into a pretty cool bar. I have a slight inkling to build another structure and use all my lab skills (plumbing, PID temperature control, etc.) to make it into a brewery. I don't think I'm a good enough brewer to "go pro," and it feels like the brewery market must be near saturation. But it's just a fun way to spend time.
5.) Online poker
I have always liked poker, and am a marginally good player. I have some statistics on limited online play:
in about 120 sittings in online play, I have a 15% return on investment over a 6 month period, but it is only 7.5% over the last 1/3 of that, or so. I think these statistics are not very robust because the sample size is low. Also, these are cash games on the order of $5-$10 buy-ins, so an average day would be sit down with $10, leave with 10.70 after an hour or something.
However, I've never invested real time in working on it, and I've definitely never employed analytics. Anyone out there able to comment on my "beginner" stats, and whether using computing to learn about optimal strategy is a window to significant improvement? EDIT Or are the robots already about to take this space over? Or is it so saturated with real quantitative thinkers and not enough weak players that it will be too competitive to make any gains? 6.) Will I actually do any of that stuff anyway?
This is my main fear of the free time route: all of these ideas are probably harder and less interesting than I think. I doubt I'll get super passionate about any of them. I might just waste all my free time and be a less interesting person than I would have been if I became a legit scientist. OTOH, maybe all that "wasted" time will just be hanging out with my kids and family more, and making them into the best human beings I can.
Thanks for reading all that! I'd love to hear any and all thoughts.
Last edited: