I was one of three researchers in a team at Yorktown Heights that did some of the best research in AI in the world, published a string of peer-reviewed papers, was the source of two commercial products, gave a paper at the AAAI IAAI conference in Stanford with for that year the 25 best AI applications in the world, personally won an award, etc.
I've taught computer science at Georgetown University and Ohio State University. My Ph.D. is in applied math.
Once in my career, just as a 'scientific programmer', in two weeks I sent a few resume copies, went on seven interviews, and got five offers.
But after my Ph.D. and AI work, I sent over 1000 resumes to Google, Microsoft, GE, FedEx, and hundreds more, got only five interviews, and no offers. I got a nice letter back from Fisher Black (as in Black-Scholes) saying that he saw no applications of applied math or AI at GS.
I ask you: Who will hire you in AI and why?
In business, hiring is because some manager has some work to do and a budget to do it. That manager believes that they know nearly all that is needed to do the work and otherwise would not be betting his career on the work. Thus, the manager is not hiring high technical expertise he doesn't have. Instead the manager is, as on a factory floor 100 years ago, hiring labor to add 'muscle' to his work.
In particular, unless the manager knows AI, he won't be hiring for AI. And there is at most only a tiny chance that the manager knows AI and even less chance that his project will depend on AI.
Moreover, the manager does not want competition from below and does not want his project 'disrupted' from below so really doesn't want technical expertise above what is needed just to get his project done.
Net, if you know some AI and want to use it in business, then find an application and start your own business. Then, since you know AI, you won't have to hire anyone in AI either.
What I've said here for AI holds for essentially all advanced academic topics.
Sorry 'bout that.
So, since this thread was about 'job placement' for students who did well in an AI class, I posted about my experience getting hired where part of my background was some expertise in AI.
Net, I have to conclude that, in getting a "job", expertise in AI will be from very rarely helpful to often a serious disqualification as in "It appears that you are overqualified for our position and would not be happy in it".
You are "overqualified". People will read your resume and assume that you will be unhappy with a regular software engineering job, and therefore not hire you for such a position.
Like it or not, you'd need to find an AI job where people needed your particular skills. (Or something closely related.)
After my PHD I got a postdoc in a 3 year EU project just finished. Now I am craving to get out of "academia" and get into software development again. The problem is that I would consider my development skills as a "junion" or "mid level" developper but without hardcore expertise in a technology.
And the worst problem is that as you say, a lot of companies that see my Resume see "PhD" and think "overqualified".
Recently I tried appliying to a group some company that is doing Machine learning with the hope that they will see a PhD as a "feature" and not a bug.
At the time, 1995-2005 (see my post below) AI was not much on the radar of companies. It would, could, and should have been but was not.
But asking that a company need "particular skills" that are a bit advanced is, as I explained, fundamentally something of a long shot.
Net, if someone has some advanced expertise and sees an application, then they should just start a business and there be CEO-CTO-CIO, and Chief Scientist along with chief floor sweeper until they get funding and/or revenue and can hire people.
Edit: for example some opportunities here http://www.kdnuggets.com/jobs/
Elsewhere on this thread I've outlined how business 'handles' new technology.
So, yes, maybe a big credit card company gets up on their hind legs about credit card fraud and wants to attack the problem with AI, big data, machine learning, optimization, etc. So they set up a group.
First problem: Some high executive looks at the group, doesn't see what he expects and respects, gets a headache, and kills the group.
Second problem: The group gets some really nice research done, writes a technical paper, and gives an executive briefing; some high executive gets a headache, concludes that the group is just engaging in 'theoretical nonsense', and kills the group.
Third problem: The group is making good progress, has some running software with good estimates that credit card fraud losses will go down by 75%, thus giving an ROI off the tops of the charts, with essentially no effect on non-fraud (false alarm) credit card operations. Some executives elsewhere in the company start to feel the heat of internal competition and work to shutdown the project.
Fourth problem: The project is rolled out in production and is fully successful. There is no more need for the group, and it is disbanded with everyone fired. The head of the group has his house foreclosed, and his wife leaves him. He sends 1000 resume copies and joins his brother's business mowing grass.
Lesson: Being a technical employee in a big company where 95% of the people and nearly all the executives are non-technical sucks, i.e., is a career long walk on a short pier.
Net, big, old companies are nearly always just unable to work effectively with new ideas. Exceptions are possible but rare in practice.
Larger lesson: Start a successful company and either run it and pass it down in the family or sell it for $50 million, $500 million, whatever. Don't be the factory floor worker. Instead start a company and be the CEO. Use special technical expertise as the advantage.
Can plug together one heck of a Web server computer for $1000 in parts. Heck, just in starting a pizza shop, one pizza oven might cost more than $1000. The pizza shop needs to pay attention to a long list of regulations, but a Web server doesn't,
One of the good things about the US is the relative ease of starting a new business. If want to use technical abilities to start a company and sell it for $50 million, then the US is the place to be. Thank the big, old companies because they are the ones willing to spend $50 million buying a company based on less than 100 KLOC!
The main issues were (1) I am a US citizen and, thus, not on an H1B and (2) I was over 50.
Generally a subordinate is supposed to be younger.
Near the end of the visit, a nice girl in their HR office walked me to a bulletin board they had with their legal announcement of their job openings and their claim that they had to hire H1Bs because no qualified US citizens were available. We didn't say anything to each other, but the scam was clear, Hope they didn't fire her.
A friend, who worked with me at Yorktown Heights, recently went on an interview for a programmer slot. Apparently all the programmers were on H1Bs from Taiwan, India, and Russia. My friend didn't get hired. He's terrific at C, C++, C#, Visual Basic, .NET, FoxPro, T-SQL, and system management and administration of Windows Server, SQL Server, and Exchange. And he's an expert in AI with an applied math Ph.D. from one of the world's best math departments. Yup, guess those guys from India, etc. were 'better qualified'.
The early part of my career was greatly helped by the Cold War and the Space Race. That's why in two weeks I could go on seven interviews and get five offers. Yes, that was near DC. At one time I was making six times what a new Camaro cost. I still have the Camaro!
But as has been documented, about then some executives from industry and government got together to see if they could change that 'horrible' situation. The result was that the NSF set up a team of economists that did some supply-demand calculations and estimated how many more 'tech workers' would be needed to 'solve the problem'.
Then to get the 'tech workers', the NSF wrote into academic research grant contracts that so many students had to be supported. And, hint, hint, hint, such students are available from Taiwan and India.
So, for some years freshman calculus classes were taught by graduate students with good understanding of Chinese but poor understanding of English. Of course, if are going to study a subject in English but don't know much English, then about the easiest subject to study is math since the vocabulary is very small and the terms are very well defined.
Soon US citizens in college walked into computer science classes and, on the first day, saw only a minority of US citizens, sensed something wrong, and walked out. So for some years during rapid growth of computing in the US, academic computer science was very short on US citizens.
Really, the H1B program was designed and intended to flood the US labor market for tech workers, and basically the program worked.
Congress was getting pushed from two sides on the H1B situation. But 9/11 provided an excuse to throttle all immigration (except from Mexico!), and the permitted H1B slots were shrunk.
Now some tech industry executives are back at claims that the US needs more immigrant entrepreneurs to get 'skills' in 'short supply' in the US. So, amazing situation: In the US families commonly have one heck of a time paying for college. Even in the US, a good computer, printer, Internet connection, work space, etc. for learning computing is somewhat expensive. A lot of bandwidth to US servers is needed for downloads. Yet somehow in countries with average family incomes 10% of those in the US and 10,000 miles farther away from US servers people are 'better trained' in 'technical skills'. Amazing.
No, it's just an old story: Economic activity needs land, labor, raw materials, capital, etc. Anyone with one of these likes to believe that their part is the most valuable. So, the people in the US with the capital, and who never wrote 100 lines of software or invented an algorithm, tend to believe that they have the most brains, the most valuable part, and should have the most power and that labor should be like workers on a factory floor 100 years ago. They wrap themselves in notions like "The US is a nation of immigrants". Yes, Mayor Bloomberg, you are one of those people.
To me these are now very old issues and there are some larger issues:
First, Moore's law and related 'laws' for other hardware have been charging along so fast that what can be assembled for a development computer or a first server for $1000 in parts is astounding.
Second, common US Internet bandwidth is beyond belief, even for a server. Just do some arithmetic assuming a Web page that sends for 200,000 bits, with three ads, with some reasonable 'charge per thousand ads displayed' (CPM), and an Internet connection with 15 Mbps upload bandwidth for less than $100 a month, assume half fill that bandwidth 24 x 7, and estimate the monthly revenue. THen can join the supercharged Corvette of the month club or the 50 foot yacht of the year club. Multiply it out and see.
Third, US technical graduate education still totally knocks the socks off nearly all the rest of the world. If have some such education and some research and also a good application, then get a computer, type in the code, go live on the Internet, get users, ads, ad revenue, a Corvette and a passenger about 5-4, 110 pounds, good figure, natural blond, cute, sweet, majored in art history, good at cooking, sewing, playing piano or violin, singing, wants to be a wife and mommy, ...!
Yes, I've published in mathematical statistics, that is, the more serious version of 'machine learning'. And I've published in optimization, that is, the more serious version of 'planning' in AI. And my Ph.D. research was in stochastic optimal control, that is, the more serious version of the AI 'planning over time under uncertainty'. And I've done some applied math research for my project. I recommend this path instead of 'computer science'.
That is, if going to grad school, I believe that there are some serious advantages in a carefully selected collection of topics in applied math instead of 'computer science'. Start with an undergraduate major in pure math. Sure, somewhere learn to write some code and then get, say, three hours of lectures on 'algorithms and data structures'. In graduate school, take seriously measure theory, functional analysis, probability based on these two, stochastic processes, optimization, and mathematical statistics, at least. My guess is that you will have the best tools for the future of computing and 'information technology' entrepreneurship and won't have much competition from outside the US or even inside. And those math classes are NOT crowded!
I need to get back to it!
Yes, I know that Google, Microsoft, etc. should be using AI for ad targeting, search ranking, scam and fraud detection, etc. Maybe they are.
Still from the tech news it does appear that the best way to get paid for such work from such companies is to do a startup with such work and then just sell your company to one of those companies.
I still believe that the ability of US big business, with its many 'traditions', will have one heck of a tough time running projects with anything advanced technically and not already common in the company. In simplest terms, the existing middle and upper management has essentially no ability to work productively with anything new.
Indeed, broadly the leading US research universities are terrific at working, for as far as they go, with new technical ideas, and business is just AWFUL at it. So, super tough to get a business to buy a technical idea, but they will buy a business based on a technical idea.
Oh there is a history: Off and on for decades business has tried to make statistics, optimization, AI, etc. work. The pattern is some new buzz words, some hype, some projects, some project failures, and then all's quiet again. Thing is, of course, eventually the ideas have to work. Business doesn't know how to make such ideas work. Parts of US research, in or close to academics, is from a little better at making ideas work to much better. E.g., look up the video of Eric Lander's lecture at Princeton 'Secrets of the Genome' or some such and see what he made work with new ideas and new hardware.
In simple terms, 'business' gets a 'business model' that is working and then hires 'line managers' to 'manage' the execution of this business model. Typically when the model dies, the business dies. Doing well with work that is new and advanced within the on-going business is not common. Yes, one of the reasons is part of the tax code -- it can look better for a company's financial condition just to buy a business for, say, $50 million, than to develop in-house.
The good news is, this is a GREAT time for information technology startup entrepreneurs. Don't struggle against your problems; instead, pursue your opportunities.
Back to it.
I'm at a 4-person startup and AI is the core of what we do. My e-mail is on my profile if you (or others) are interested.
You sound like a qualified candidate, if only you'd sent your resume our way.
Finally I gave up on getting hired and decided to start a business.
I'm now deep into that effort and at least for now should continue it.