20 ms·

I have a PhD in math but I often dream about how nice it would be to get paid to write simulations for big shot mathematicians. Usually this work goes to underg

I have a PhD in math but I often dream about how nice it would be to get paid to write simulations for big shot mathematicians. Usually this work goes to undergrad/grad students. Maybe it’s just not worth paying for professional code.

Ma8ee 6y ago
It’s often hard to get money to fund anything else than PhD students or post docs.
- Kalium 6y ago
  I suspect it may also be hard to swallow the idea that a professional engineer's skills are needed when such a person costs more than the researcher's salary.
  - Ma8ee 6y ago
    I actually don’t think that is the problem. In addition to the shortage of funds, there’s often a culture of doing most things yourself. When I was a PhD student, I did everything from vacuum systems, electronics, high current equipment, plumbing for cooling water, setting up a real time control system for all the equipment, data collection and analysis, and programmed quantum Monte Carlo simulation that was run on a high performance cluster (on top of writing papers and a thesis). Each of these tasks would have probably been done better by someone specialising in it, but there was certainly no money for that. (I had almost forgotten how much fun I had.)
    - SiempreViernes 6y ago
      Doing it yourself is certainly very important to be sure you get at least some understanding of what's going on, and often you are doing things nobody else knows how to do anyway. But it sure would have been nice to have some professional support programmer around for advice on design and help out with the gnarlier bits.
      - Ma8ee 6y ago
        Programming was actually what I did best. I had been programming since I was eleven and also took some computer science courses at the university. In addition, the kind of programming that most physicists do isn’t really helped by the skill many so called engineers have. It’s not about organising large multitude of systems and their interaction (which is usually the hard parts today) as much as knowing things about efficient numerical packages and making sure that the computer’s finite precision don’t overwhelm your statistical uncertainties. And of course, to understand the model, with its limitations, of the system we are trying to simulate and efficient algorithms for that.
- chriskanan 6y ago
  Yep, and even then they don't get paid super well. A typical "small" NSF grant is for $500K over 3 years, and the university takes 45-65% of that, so with one award a PI has ~$83K to spread on their annual summer salary and ~2 PhD students at a stipend of around $30K per year each (with the remainder on conferences, publication fees, equipment, etc.).

Ma8ee 6y ago

It’s often hard to get money to fund anything else than PhD students or post docs.

Kalium 6y ago
I suspect it may also be hard to swallow the idea that a professional engineer's skills are needed when such a person costs more than the researcher's salary.
- Ma8ee 6y ago
  I actually don’t think that is the problem. In addition to the shortage of funds, there’s often a culture of doing most things yourself. When I was a PhD student, I did everything from vacuum systems, electronics, high current equipment, plumbing for cooling water, setting up a real time control system for all the equipment, data collection and analysis, and programmed quantum Monte Carlo simulation that was run on a high performance cluster (on top of writing papers and a thesis). Each of these tasks would have probably been done better by someone specialising in it, but there was certainly no money for that. (I had almost forgotten how much fun I had.)
  - SiempreViernes 6y ago
    Doing it yourself is certainly very important to be sure you get at least some understanding of what's going on, and often you are doing things nobody else knows how to do anyway. But it sure would have been nice to have some professional support programmer around for advice on design and help out with the gnarlier bits.
    - Ma8ee 6y ago
      Programming was actually what I did best. I had been programming since I was eleven and also took some computer science courses at the university. In addition, the kind of programming that most physicists do isn’t really helped by the skill many so called engineers have. It’s not about organising large multitude of systems and their interaction (which is usually the hard parts today) as much as knowing things about efficient numerical packages and making sure that the computer’s finite precision don’t overwhelm your statistical uncertainties. And of course, to understand the model, with its limitations, of the system we are trying to simulate and efficient algorithms for that.
chriskanan 6y ago
Yep, and even then they don't get paid super well. A typical "small" NSF grant is for $500K over 3 years, and the university takes 45-65% of that, so with one award a PI has ~$83K to spread on their annual summer salary and ~2 PhD students at a stipend of around $30K per year each (with the remainder on conferences, publication fees, equipment, etc.).