6 July 2009 · 1 min
Goodreads is a Social Web site about books. They need a recommender system. Thus, they issued a challenge: design their recommendation engine, and they will read your resume. I suppose this is the poor man’s version of the Netflix challenge.
There is nothing wrong with the challenge, but I wish the current Goodreads engineers would publish their own recommender system. How hard can it be to do item-based analysis?
Update: Students at Stanford built and tested a recommender system for Goodreads.
Daniel Lemire, "Design a recommender system: they read your resume," in Daniel Lemire's blog, July 6, 2009, https://lemire.me/blog/2009/07/06/design-a-recommender-system-they-read-your-resume/.
[BibTeX]
I read that more as an interview question – and I believe such a question would be effective.
ITA has some crazy good puzzles; Google’s also used them.
It isn’t a terribly difficult challenge; anyone that’s picked up Programming Collective Intelligence should be able to give a decent answer. Finding those few people that do read books, however, is actually quite difficult.
Sure. If I were looking for job in their company, I’d answer them. Nothing wrong with the challenge.
This sounds like a very clever commercial proposal. They get the solution free and could amalgamate the best of several ideas submitted. Quite an education for those who loose out.