Jesse Farmer, a software engineer, writes about programming, interview questions, Facebook, startups, and metrics.
HN links it for its Erlang and graph theory tutorials, with commenters noting broken links and a copied version of one article.
San Francisco, CA
Hi HN! I'm Jesse, one of the founders of Everlane.
http://www.everlane.com
We're trying to re-imagine retail online and make it easy to find products that match your taste and style. Right now we're focused on mens fashion, but our ambitions are much larger.
pg did a good job of explaining the opportunity, here: http://ycombinator.com/rfs2.html
We're building a members-only club where each month we produce and release a new collection of mens accessories and apparel. Yes, we're selling our own products, directly to consumers.
This lets us offer a tie you'd buy for $80 at Nordstrom's for half that. To produce in-style products we have a team of curators, run the largest men's fashion Tumblr, and are generally data-savvy people.
You can read my (somewhat out of date) blog, here: http://20bits.com
We're well-funded by the most stylish investors around.
We offer competitive salary, equity, and benefits. Our office is in Union Square in San Francisco, CA.
If you're a hacker, love data, and/or love style shoot me an email at jesse@everlane.com and get in touch.
Also, if you're up to the challenge, here is a little exercise for you: http://static.everlane.com/everlane_exercise.pdf
A point to add to the discussion from the metrics perspective: the net promoter score.
Net promoter score is an easy tool you can use to determine trends in your users' experience. It can be used as a leading indicator of the 'N' part of your viral coefficient. eries goes in to detail in "Net Promoter Score: an operational tool to measure customer satisfaction" (http://startuplessonslearned.blogspot.com/2008/11/net-promoter-score-operational-tool-to.html).
Additional reading from HN members:
jfarmer's "Three myths of Viral Growth" (http://20bits.com/articles/three-myths-of-viral-growth/)
I think there was an error in his analysis of the naive solution for finding loops in linked lists (at http://20bits.com/2008/04/17/interview-questions-loops-in-linked-lists/)
I don't know Ruby, so correct me if I'm mistaken, but I believe the naive solution presented runs in O(n) space and O(n^2) time (not O(n) time, as the author suggests).
If you used a hash table to keep track of nodes seen so far, you could do O(n) time and space.
The tortoise and the hare solution is quite elegant though.
An interesting supplemental reading is the "two bowling balls" interview puzzle -- http://20bits.com/articles/interview-questions-two-bowling-balls/
It's essentially a 2 level skip list (but could be generalized where you get one extra level per bowling ball), and brings out some basic calculus to optimize lookup performance even further.
"... An important tip, not mentioned in the article: 'blog about Web 2.0 aimed first and foremost at programmers, computer scientists, and mathematicians' ..."
Looked at the site, interesting ~ http://20bits.com/2007/04/27/facebook-job-puzzles-prime-bits/ Maybe they are paid to code, but not to think "why"?
His direct links from part 3 to parts 1 and 2 are broken. However if you click on "More Articles" (http://20bits.com/category/articles/) on his homepage the links there work fine.
http://20bits.com/articles/graph-theory-part-i-introduction/
http://20bits.com/articles/graph-theory-part-ii-linear-algebra/