I think quality begets quality, too. I once had a phone screen interview with Backtype before they were bought by Twitter. I was (and perhaps still am) extremely unqualified for the role, but the interview process was interesting and enjoyable. At the end, after doing a small at-home project for them, they turned me down. To thank me for their time, they gave me a 25$ Amazon gift card, which is still the best rejection gift I've received to this date. I worked for Amazon at the time, and you basically get $100 in Amazon credit per year even as a full-time employee, so the irony of getting a quarter of that for failing an interview was not lost on me.
I used to work at Amazon too. The 100$ year credit is actually up to 10% of your purchases, so you need to buy at least 1,000$ of stuff in order to get that 100$ credit. Given than Amazon earns money from selling stuff, they probably not even lose money with that 100$ credit. It's simply a ridicolous perk. I hated these things at Amazon. Others were interesting and nice.
More than anything this article convinces me that there is a bubble of Data Scientists (speaking as someone who as employed as a DS and quit one month ago to build his startup).
Well, my previous job was more data pulling involved than statistical modelling. It could have been just me, but the company had more than 30 DSs and I think only two of them did anything advanced. I read on HN once that 95% of the DSs where in the same situation as me.
So for me it seems that Data Science is more a fad and company are hiring PhDs more for the sake of having an impressive data science team than for using advanced modelling tools.
I may be wrong of course, but Nathan mentions in his blog Netflix and Khan Academy. I guess Netflix can use data science to make recommendation system, and Khan Academy to gauge the level of the exercises they propose, and see if the progression curve of the students is right. Honestly, I don't see how this would need something particularly advanced in term of modelling.
Does that translate to: "it's easy to get a DS job now adays since no one is actually doing anything super advanced. So its unlikely you are under qualified for the job" ??
Kind of. I am under the impression that many companies have a top management that wants a strong data science department, resulting lots of demand for high profile PhDs, without having actually a real use for them, hence many DSs are over qualified, like I was.
I've had the opportunity to meet Jake Klamka in person, and he is one of the most charismatic people I've met. His Insight program attempts to solve a really important problem in our world. It doesn't just solve the academic-job-prospect problem, but the way to interview people isn't limited to a 4 hour or 2 day interview anymore - employers can see the demonstrated work output of these talented candidates.
I liked this article a lot. The art of approaching a business opportunity. Jake approached him like he would a potential buyer of his products. Don't force a sale on someone, get them to want your product first and eventually have them begging you for it. Very clever.
