Programming, Finance what else can a Physicist do?

[Crimson_Aether]

I'm currently a database programmer, which is ok but repetative.
What it lacks (in my experience at least) is the spirit that my degree had. I miss thinking analytically, doing integrals, using my brain to make things from scratch rather than memorising the motions of something already invented.

Does anyone have experience with/knowledge of any sort of work outside of doing a PhD which is like this that a physics graduate can do?

*Inclusion of additional requirements is fine, for instance "you'd need a masters in X" or other qualification, etc, the intent of the question is to establish general sectors of employment that might fit the above criteria or deliver a similar experience rather than a specific job or company, though these are welcome too.

 

[K Ramachandran]

Insurance companies need mathematicians for the job of underwriting. I don't know whether this is repetitive or interesting . Humana. Louisville, has a department for this.

 

[Locrian]

In health insurance, underwriters typically bridge the gap between actuaries and the actual member groups and the sales associates who serve them. Underwriters will contact groups, gather information about them, use pre-defined formulas to set rates, and monitor and maintain groups afterwards.

Underwriting has always appeared to be interesting work to me, but it’s never seemed particularly technical in nature, and certainly doesn’t require much in the way of mathematics. My guess is that underwriter’s technical proficiencies probably have a wide range, depending on which function they’re serving. Those leading or managing the underwriting team tend to have very good business skills and a great deal of knowledge and history of the business.

If you consider underwriting jobs, also consider other analytical jobs as well, such as business consulting jobs with technical roles, metrician roles, anything with “analyst” in the title and roles that support actuarial work. I’ve actually run into business consulting roles earning $80k+, which surprised me, and the work can involve a good amount of problem solving. There are also some pretty dull jobs with those titles, but once you get one, you can move to another you might prefer.

 

[MarneMath]

Data Scientist comes to mind. A masters in comp sci is generally enough to get an entry position. You wouldn't get paid as well as a Sr. Data Scientist (aka guys with PhD) but most of my day is spent reading technical papers and doing statistics, so I really can't complain.

 

[Cygnus_A]

Data Scientist comes to mind. A masters in comp sci is generally enough to get an entry position. You wouldn't get paid as well as a Sr. Data Scientist (aka guys with PhD) but most of my day is spent reading technical papers and doing statistics, so I really can't complain.

If you don't mind me asking, can you describe the kind of work that you do? What kind of statistics do you use? Are you mainly applying it to computer programs?

 

[SteamKing]

Insurance companies need mathematicians for the job of underwriting. I don't know whether this is repetitive or interesting . Humana. Louisville, has a department for this.

This is the work which actuaries do. They help gauge the risk to which insurance companies are exposed, so that premiums can be developed.

Actuaries take a lot of math, particularly statistics, and they reportedly have to take a rather comprehensive professional examination in order to become one.

 

[MarneMath]

If you don't mind me asking, can you describe the kind of work that you do? What kind of statistics do you use? Are you mainly applying it to computer programs?

I make models and predictions. My main goal is to identify problems before they become critical and attempt to save my company money by fixing them before a complaint or finding the optimal time before a complaint. These involves using various languages like R, Python, hadoop, scala, scalding, and I once even used perl :p. Oh I should mention C and kuda but I'm really into optimizing speed, but a lot of my peers are not. The work is generally high level because you'll find that taking real world data and turning it into math is actually non-trivial. It requires a good deal of computer skills, math skills and domain expertise. However, at the end of the day I find it rather rewarding and intellectually challenging.

The statistics I use is wide and varying. There's the simple correlation, means, std, p-value stuff, which is useful in understanding the information. In the model building process, I have used grubb's test to identify anomalies, mann-kendall test for trend analysis, continuous markov models for predictive fault testing, survival analysis for predictive call analysis, along with other techniques for various of task that relate to signal processing.

Beyond the statistics, there's also machine learning algorithm, which some would call 'engineered' statistical learning. So a lot of support vector machines, neural networks, affinity clustering, etc.

As for applying it to computer programs, I actually rarely write statistical techniques and turn them into programs. I have rewritten poorly written code, and I have rewrote good code into another language, but I rarely have spent time reducing the time and space complexity of an algorithm for computational purposes. There do exist Data Scientist who do this, and enjoy it, I just happen not to be one of them. Although, in my new role, I'm starting to get into as 10 Terabytes of data is becoming my new norm, I feel that learning more about using the advantage of parallel computing and map reduce paradigm would probably help me a bit.

 

[Cygnus_A]

That was very informative! Thanks!

I'm in the process of choosing a PhD topic, and I might one day want to transition from academic computational physics (say, CFD or Cond. Matter physics) to an industry computational job; although, I have a feeling that your job requires a lot more statistics in it than a physics PhD would learn right off the bat. On the other hand, I've been leisurely teaching myself some interesting stats applications (like markov models).

 

[Laura Forczyk]

I'm a nearly-PhD in physics (ABD) and I help to run a newspace company. My husband is a PhD economist and he says if I wanted to to enter the world of finance, I could be a quant (quantitative analyst) like him. My father-in-law has a master's in physics and works on sensors for the defense industry. I know of one who went into journalism. So very many options. Focus on your goal and consider the skills you need to make that happen.