mikeinterviewst

InterviewStreet (http://interviewstreet.com) - Mountain View and Bangalore

Programmers for frontend, backend, and project management roles. Our hacker team is 5 strong, so you'll be tackling a huge variety of projects. Help fix a horribly broken hiring world by creating the best platform for addictive programming problems and tutorials.

team+mv or team+blr at interviewstreet dot com

That was how we originally built it, but we found it less fun to rate in testing compared to this version. Hopefully the increased votes-per-person outweigh the loss of fidelity.

Mountain View (Remote OK for first couple of months)

InterviewStreet (YC '11) is solving technical hiring through interesting, practical problems (like Stripe's recent CTF). You'd be employee #7.

We're looking for:

* Campus Coordinator (non-technical): http://blog.interviewstreet.com/2012/02/help-interviewstreet...

* UX / UI designer: https://news.ycombinator.com/item?id=3553649

For both, send an email to [email protected]

The biology revolution is coming, both sooner and later than you think. As one indicator, the price per base pair of DNA synthesis is following a Moore's Law, (http://singularity.com/charts/page73.html), but Fredster is right in that there are other factors at play. However, you could say the same with computers, with bandwidth, cpu, and storage all improving simultaneously.

A story that our TA in EE40 (Signals in Systems) at Berkeley really stuck with me: in the early days of personal computing, when you would go to Radio Shack to buy a set of transistors, you had to individually test each one, because they had a 50% chance of not performing to spec. Biology currently is in that stage, though multiplied by another order of magnitude.

The topic that patio11 mentioned is covered pretty well in chapter 2 of Freakonomics. The actual text doesn't seem to be part of the official sample excerpt, but somebody posted the relevant portion in a forum here: http://www.nachi.org/forum/f11/exerpt-freakonomics-book-rega...

This is not an endorsement of other portions of the book as a whole.

Scientific American had a great article on this in my Pinboard: "The Secret of Raising Smart Kids: Don't Tell Them They Are" http://www.scientificamerican.com/article.cfm?id=the-secret-...

EDIT: It looks like it's behind a semi-paywall. Here's a Googled PDF: http://www.ccsf.edu/Campuses/Downtown/scientific_american.pd...

It looks like your privacy settings / a plugin you have installed doesn't allow for even obsfucated email sharing. Normally, what we get for Facebook logins is app+<random_string>@proxymail.facebook.com. This is the first time we've heard of that it's not able to be shared.

Wow, we did not expect to see this level of criticism from the comments.

A lot of the focus is on the 48 hour period, in that it's basically way too long to expect you guys to participate. We don't expect people to code all 48 hours; the decision was basically "make it a weekend long", but we are not tied in any way to that time period. Would it be much better if we shortened the time period to 24 hours? (We can't shorten it any more than that; as this is a global competition, everyone needs to have the same amount of daytime) The goal is not to have you guys burn yourselves out. The goal is to assess you all as quickly as possible.

From our perspective, we sort of viewed CodeSprint as a time saver; everyone you apply to nowadays requires some sort of technical proof. Sometimes you can get away with it if you have a github with a lot of open source projects, but not everyone has that.

With CodeSprint, you get to choose among a variety of problems, and companies with those specific problems get to find you. Because they receive the full code solutions, our goal is for CodeSprint to consolidate and replace the "do-a-technical-problem" part of the normal interview process. Except, this time around, this one set of problems is what's used for multiple companies, instead of just one.

Companies choose candidates on a per-question weighting system. Thus, if you spend only a couple of hours, but on a problem that is valued highly by a specific company, then there's a good chance that they'll still notice you.

It's free for programmers to participate.

Just to clarify a quote: “Everything will be based on performance — how fast they code, how many they get right.”

This is a slight misrepresentation of the original intent. What we meant to say is that we measure code execution speed, not how fast people type.

Which university is it? We'll manually add it in, and if we get a lot more comments about universities not being city-specific, we'll make the change.

Maybe. If so, we will tell you about it in our Post Contest blog post.

Based on the IRC chat, most of the battle is tackling our biggest test cases and finding an algorithm that fits into our memory and runtime limits.

The contest was designed for busy students in mind. We targeted this test to be somewhere between 12-20 hours long in total.

Yes, we will have a CodeSprint Post Contest Results post.

We're mainly focused on US university students, so we don't anticipate it being a big problem, but we do have company preferences in the profile (yet to go live), so students can pick and choose which companies they want. We'll get the encoding fixed.

Time-lapse was a "cool feature" requested by one of our early companies that we decided to implement on the main InterviewStreet site. If you cut and paste completed code, it won't affect your score whatsoever, and it's up to the specific company whether or not they consider being able to see your in-code thinking process as important.

A little bit of both.

There will be a hard cutoff, in which we are only considering score.

With that said, in the profile (which hasn't launched yet), we let you fill out fields / github accounts to further flesh out who you are. Companies are given a complete profile of students, and then give us their preferences on who they'd like to interview.

We liken ourselves to the Common App. Consider the contest portion your SAT score, and your profile page everything else.

To answer your first point: Our contest will cover different areas of problems, and certain companies will only be available for application if specific questions types are answered.

One of the main appeals for CodeSprint is that you don't have to work on the same puzzle twice. A lot of companies now have a FizzBuzz-style filter, and we're trying to make it so you have to FizzBuzz only once.

We're still debating internally about how to match startups with top scorers. We want to include both applicant and company preferences.

Ultimately, I think we're going to have to wait until we see the distribution of scores before we make a decision. Secretly, I'm hoping there will be incredibly obvious cutoffs for bracketing, as opposed to some of the Dutch Auction ideas being thrown around.

With that said, if you're in the top 10, and want to interview all of the companies, I'm fairly sure we can arrange that.