[o_safadinho]

I read the article correctly. They don't call it homework in the article and instead decide to use the term "data challenge". You can call it whatever you want it is the same thing. It is also split in to four parts and takes place over several days. How did I interpret this incorrectly?

I'm specifically talking about other technical fields that are highly mathematical and that have a lot in common with Computer Science.The fields that I specifically mentioned were Statistics, Economics, Actuarial Science and Operations Research. I mentioned all of the fields that I'm talking about in the original post and in sever of my replies. You are harping on something trying to correct a point that I'm not even trying to make and your response actually helps to further prove my point. You could be correct about the Chemistry thing. I didn't study Chemistry and I don't want to be a Chemist and I didn't mention the subject anywhere in my OP, so I don't really care about the interview process for hiring a Chemist.

For example, with Economics (one of the fields that I actually mentioned), presenting your "Job Paper" is actually part of the interview process. You make your presentation and everybody that has arranged an interview with you can come by and hear your talk. They can read your paper before hand so they have time to think questions well ahead of time. The American Economics Association provides one platform for all interested parties to come together and hash things out instead of me having to try and go to 10 separate interviews.

With Actuaries (another field that I actually mentioned) there is no need for you to quiz me to see if I actually know how a Poisson Process o GLMs work because that is taken care of by the exam process. If somebody passed the Statistical Models exam then you know that they at least meet the industry agreed upon minimum competency level for knowledge of specific Statistical Models.

[dalke]

I apologize. It was I who was mistaken. I overlooked that the basic question takes place at home.

This is, I think, equivalent to the FizzBuzz problem that was going around a few years ago. I can see how it can be called homework. I assume it's a screen because a lot of people apply who have no clue on what data science is actually about. If my assumption is correct, the it should take no more than 15 minutes or so, since it's supposed to be "easy."

You write now "and takes place over several days". Earlier you wrote "multi-day long technical screens with multiple homework assignments." These are two different, though related things. My response was to your first description, not your updated description.

Here are the differences I see. 1) the AirBNB text says "We send a few datasets to them and ask a basic question; the exercise should be easy for anyone who has experience." The use of 'the exercise' looks like it's one homework assignment, while you say it's "multiple homework assignments". This is a minor point, but still a difference.

2) the AirBNB text says that the in-house challenge is a day long ("They then have the day ... At the end of the day"). That is the main technical screen. The remaining day is a non-technical screen. I agree that it's a multi-day long process, but only one of those days is for technical screening. This is why I disagree with your initial characterization of "multi-day long technical screens", though I agree with your new characterization of "takes place over several days".

> "I'm specifically talking about other technical fields that are highly mathematical and that have a lot in common with Computer Science. ... so I don't really care about the interview process for hiring a Chemist."

I was addressing your earlier statement "What is it about applying the same techniques to tech that makes them special and some how harder to measure?"

In comparing your two sentences, I'm confused. You first imply that CS is "tech" and somehow special, and different from statistics and the other fields. But now you say they have a lot in common? What does "special" even mean then?

Most of CS, by the way, has very little mathematics in it. Surely less than chemistry does. Why do you think CS is "highly mathematical" when most CS departments require only basic calculus, discrete math 2, and perhaps an intro. statistics course? (Some specialized subfields of course require more mathematics.)

I'm "harping" on the topic because you think CS is somehow special, when my experience from physics, chemistry, and bioinformatics says that the CS hiring method is pretty standard.

I believe the hiring process in mathematics is also similar to the science fields I just listed, and different from what you have described.

For example, http://www.siam.org/pdf/news/1847.pdf describes the process for "recent PhDs who are looking for positions at undergraduate-focused schools". It includes a strong teaching component, so the interviewees are expected to put together a talk with "the daunting task of showing that you can connect with an audience of undergraduates and non-specialists while simultaneously making it clear that you’re a serious mathematician, knowledgeable about your area and with substantial work to your credit."

http://blogs.siam.org/how-and-why-to-ask-good-questions-duri... says "In the U.S., the typical interview for a PhD-level position lasts a full day or more, including a one-hour seminar and one-on-one meetings with researchers and/or faculty as well as managers/administrators." This is also different than your description of "You make your presentation and everybody that has arranged an interview with you can come by and hear your talk."

Am I incorrect? Is there a centralized system, perhaps at SIAM, for mathematics papers and job finding which is similar to what you describe for economics?

[o_safadinho]

I'm not sure about Mathematics, but for Statistics the American Statistical Association hold the 'Joint Statistical Meeting' every year. It is part continuing education classes, research presentations and career fair. Every year there are dozens of companies, universities and national research labs interviewing there. They also do a regular salary survey so you know what types of salaries are typical for Masters/PhD holders in different parts of the country and different industries.

If you plan working in predictive modeling at an insurance company you will have to get credentialed by one of the two actuarial societies. The SOA just added more exams to their current schedule to cover various predictive modeling topics and the CAS decided to make the predictive modeling portion a separate credential. But this way, you do everything once and it is valid across most English speaking countries. It doesn't matter whether you want to work at AllState, ESurance or Liberty Mutual. They just pull your record with one of the Actuarial societies.

I was saying that depending on what you specialize in they do have a lot in common. Given that you have similar specialties, why are economists and actuaries and statisticians able to judge a persons competence in some sort of standardized way but a tech company trying to hire someone in Machine Learning or Data Mining isn't able to do the same?

[dalke]

All fields have big yearly meetings as you describe. And people interviewing at them.

Including, as I mentioned, ones organized by the ACM and IEEE. Here's the ACM's stepping-off point for graduating students: https://www.acm.org/education/resources-for-grads . They have a twice monthly publication called CareerNews. The ACM links to the many resources available for getting an idea of pay scales in different parts of the country.

At this point, I don't know what to say. Yes, a few fields have credentialed requirements, ranging from haircutter to certified public accountant to lawyer to surgeon to engineer.

Most do not.

For example, if you plan working in predictive modeling at drug discovery company, no such credentials exist. Moreover, a credential from an actuarial society will almost worthless, and surely less important than a MS in biostatistics.

You keep coming back to: "why are economists and actuaries and statisticians able to judge a persons competence in some sort of standardized way".

Shrug I know nothing about those fields. For all I know, it's because there is a well-known set of laws and regulations they have to follow, and generally accepted practices, so most of the need for competence is to ensure that people know those laws and practices.

That's not the case for most fields. Including computer science, in all of its forms and incarnations.