6 ms·

I don't know, I haven't had the same experience as you. I did a lot of integer programming in grad school. While you're right that integer programming doesn't

I don't know, I haven't had the same experience as you. I did a lot of integer programming in grad school. While you're right that integer programming doesn't scale well enough to solve many real world problems [1], OR and and optimization skills have helped me have a nice career. I did real word scheduling for a supply chain company (via constraint programming and GA's), worked at a quant hedge fund (where I formulated lots of classical portfolio optimization problems), worked at an energy analytics company (lots of QP's). Recently I've been working in Machine Learning, which has tons of optimization problems. You're just trying to find the optimal predictions, at the end of the day. So I've been making a nice living with my OR skills in many ways. I'm a proud INFORMS member ;-) [1] Edit: and even that is changing. The 2016 Mixed Integer Programming Workshop [https://sites.google.com/site/mipworkshop2016/speakers https://sites.google.com/site/mipworkshop2016/speakers] seemed to have a lot of interesting work. Take this paper by Dimitris Bertsimas on using integer programming for variable selection in ML: http://www.mit.edu/~dbertsim/papers/Machine%20Learning%20under%20a%20Modern%20Optimization%20Lens/Best%20Subset%20Selection%20via%20a%20Modern%20Optimization%20Lens.pdf http://www.mit.edu/~dbertsim/papers/Machine%20Learning%20und...

graycat 10y ago
I had an excellent background in optimization in grad school -- world class research university Ph.D. -- and a good practical introduction at FedEx, on various DoD problems, and on several commercial problems. I tried to have a career much like you had. E.g., at one time I sent a letter and resume to Fisher Black at Goldman Sachs and got back a nice answer that he saw no role for optimization on Wall Street. I sent resumes, 1000+, all around, and got back next to nothing. Heck, in comparison, early in my career, just in scientific software, in one two week period I sent a few resume copies, went on seven interviews, and got five good offers. After my Ph.D., I was essentially 100% permanently, absolutely, positively unemployable at anything where I mentioned I had a Ph.D. I concluded that, to make money out of such applied math, do a startup where the applied math, likely original, advanced, and powerful, was the crucial core secret sauce, proprietary intellectual property, technological barrier to entry, and let the customers/users see only the results that they will like and without any hint that there's some math at the core. At this point I have an apparently quite good practical problem identified, the original applied math derived (theorems, proofs, in TeX), the business model developed, the software and server farm architecture and the relational database schema defined, and all the software I need to go live written and running. The production part of the software is about 20,000 programming language statements in 80,000 lines of typing (with comments and blank lines). All the software appears to run as intended. The timings are quite good. Currently I'm alpha testing the software and looking to go live, get publicity, users, ads, and revenue ASAP. But, for working, as a consultant or an employee, in optimization for others who are not experts in optimization -- I gave up. My view is that it is irresponsible and unethical to suggest that a young person study optimization.
- GFK_of_xmaspast 10y ago
  > After my Ph.D., I was essentially 100% permanently, absolutely, positively unemployable at anything where I mentioned I had a Ph.D. My last three jobs, and last five job offers, only materialized because I have a PhD in math.

graycat 10y ago

I had an excellent background in optimization in grad school -- world class research university Ph.D. -- and a good practical introduction at FedEx, on various DoD problems, and on several commercial problems. I tried to have a career much like you had. E.g., at one time I sent a letter and resume to Fisher Black at Goldman Sachs and got back a nice answer that he saw no role for optimization on Wall Street. I sent resumes, 1000+, all around, and got back next to nothing. Heck, in comparison, early in my career, just in scientific software, in one two week period I sent a few resume copies, went on seven interviews, and got five good offers. After my Ph.D., I was essentially 100% permanently, absolutely, positively unemployable at anything where I mentioned I had a Ph.D. I concluded that, to make money out of such applied math, do a startup where the applied math, likely original, advanced, and powerful, was the crucial core secret sauce, proprietary intellectual property, technological barrier to entry, and let the customers/users see only the results that they will like and without any hint that there's some math at the core. At this point I have an apparently quite good practical problem identified, the original applied math derived (theorems, proofs, in TeX), the business model developed, the software and server farm architecture and the relational database schema defined, and all the software I need to go live written and running. The production part of the software is about 20,000 programming language statements in 80,000 lines of typing (with comments and blank lines). All the software appears to run as intended. The timings are quite good. Currently I'm alpha testing the software and looking to go live, get publicity, users, ads, and revenue ASAP. But, for working, as a consultant or an employee, in optimization for others who are not experts in optimization -- I gave up. My view is that it is irresponsible and unethical to suggest that a young person study optimization.

GFK_of_xmaspast 10y ago
> After my Ph.D., I was essentially 100% permanently, absolutely, positively unemployable at anything where I mentioned I had a Ph.D. My last three jobs, and last five job offers, only materialized because I have a PhD in math.