[raverbashing]

However, with the abysmal standards of hiring, I'm sure a lot of companies would pass on very good candidates because they won't write FizzBuzz on the board for you, or companies would pass on Peter Norvig because his code is not Pep8 compliant

[SilasX]

But there would still need to be a fizzbuzz style filter for AI jobs to keep out the people who overstate their abilities. Is there a good test for basic AI job capability?

[raverbashing]

You just interview people.

Ask them to explain what a SVM is. Ask them to explain how training a linear perceptron works. This kind of stuff.

[CHY872]

Sketchy, but in the other direction. In high school I did a project that involved a support vector machine, and over the course of this I learnt how they worked to a reasonable level of detail.

Two years later, I interviewed at a large SV company, and they asked what projects I'd been working on. I gave a description of this project (2 years prior) and my explanation was phenomenal; in order to understand SVMs two years prior (without the 'necessary' mathematical background) I needed to develop all of the intuition (up to 'is kinda a high pass filter' etc) (which you might not ordinarily do at a time-pressured university course).

The interviewer was correspondingly impressed, and the SV company gave me an internship almost directly off the back of this interview.

The kicker is that at this point, I had a rudimentary knowledge of linear algebra, and absolutely no knowledge of any other machine learning; I had no business interning in their data science team.

My point being that even as a first pass, the bookwork questions don't work fantastically. FizzBuzz is no better, but a data science alternative would have weeded me out pretty quickly.

[rjaco31]

You're pretty harsh with yourself. You showed that you were interested in the topic and that you could learn it by yourself. I don't think anyone expect an intern to already master the subject he'll be working on, picking a smart motivated person is usually what you're looking for.

[CHY872]

Ah, but if you combine it with the fact that when I did study those things I didn't really enjoy them, you can probably see why I look back on the experience with healthy scepticism :P

[grayclhn]

The question I'm more interested in: when you worked there as an intern, did you suck?

(If you didn't, maybe the interview worked...)

Apologies in advance for the bluntness, and don't feel like you need to answer.

[SilasX]

That tests book knowledge, not actual understanding of how to think about AI, and leaves itself open to interviewer bias, the very things fizzbuzz avoids.

[raverbashing]

The idea that FizzBuzz has no interviewer bias is laughable at best.

I can guarantee some will pick and reject people because they did things in a way they didn't like

[sigsergv]

No it does not. Without practical experience one just cannot discuss these things freely, and pure-book knowledge is clearly visible instantly.

[jacquesm]

The Chinese room?

[eru]

I'd hire a fast-enough Chinese room.

[rifung]

Not sure if you're being serious but are you sure they would even care whether these kinds of candidates could program? I am skeptical they hired them to code; I imagine they mostly spend their time doing research and then have the SDEs implement things.

[jsweojtj]

It happens all the time, and it comes from a mix of things.

One way it happens is that you get a PhD in astrophysics with years of data analysis experience in for a data science job. Have a software engineer interview her and he might find that she doesn't know a number of basic computer sciences concepts [traversing a linked list, tail recursion, implement breadth-first-search]. His knowledge background says these basic ideas are fundamental, there are therefore serious questions about the technical ability of the interviewee.

[raverbashing]

This is a good example. At the same time this CS interviewer may not know what's the second central moment of a probability density function.

[eli_gottlieb]

Uhhh... the variance?

[kriro]

I have no idea how hiring in this field works but I'd expect a data scientist to have a pretty good (algorithmic) programming background in this day and age. You pretty much have to "play with the data" and get a good intuition for it when it comes to gigantic data sets and programming is how you accomplish that.

Or in other words...I'd be skeptical if a candidate hadn't learned programming on their own even if it wasn't required because it's pretty much impossible to get any practical experience otherwise.

[mhuffman]

I think often then have experience in programming, but in languages that don't map well to the "real world" of programming -- R, octave, matlab, etc. Those languages are also usually loaded with very helpful libraries to avoid having to do any nitty-gritty programming.