ithayer

There are several clojure shops and a bunch of clojure devs lurking around, hit me up, email is at gmail, prefixed with hn username.

this is really great advice. did crossfit for a bit and stopped, but some of the mobility stretches using a lacrosse ball and band have changed my life. (warm up first)

Avant (fmrly DebtEye YC S2011) | Los Angeles, Chicago | Backend, Frontend, Mobile, Analytics/Machine Learning | Full-Time

Ruby, R, Python, Redshift, Postgres, Scala

Avant Credit Corp ('Avant') is a fast-growing startup that offers tremendous advancement opportunities to its employees. Since its launch in December 2012, Avant, a consumer finance company, has changed the way customers borrow money -- providing them with a new and unique online experience. Avant uses state-of-the-art analytics methods in order to offer its lowest possible interest rates to the widest range of customers. Since inception, Avant has secured over $300 million in equity financing and over $700 million in debt financing to continue to grow at an exponential pace. Previously, Avant's founders created the leading international online short-term lender currently employing over 1,000 individuals and generating over $700 million in annual revenue. Avant’s founders have also launched an innovative real estate investment trust that provides affordable housing options in Chicago, Indianapolis and Baltimore with over 10,000 apartments owned and managed.

Please apply at https://www.avant.com/jobs

ReadyForZero (YC S2010) | Los Angeles, CA | Backend, Frontend, Mobile, Analytics/Machine Learning | Full-Time

Clojure, ClojureScript, R

ReadyForZero's award winning product and blog helps consumers get out of debt. We've helped hundreds of thousands of people pay off >$200M in debt. We were recently acquired by Avant, a Chicago-based online lender to offer our product to their borrowers and are hiring aggressively in Los Angeles.

Please reach out at jobs@readyforzero.com

https://www.readyforzero.com

Author here, thanks for the comment -- since that seemed to have been lost, I'm using "fundamental" because transducers let you describe your logic so that it that can be applied over sequences and non-sequences in a way that you cannot by just applying composed logic functions through existing clojure machinery.

EDIT: s/composed functions/composed logic functions/

Probably way beyond the scope of that project, but actually this is a pretty interesting and well-studied problem called segmentation, where you train a system to insert the correct breaks. This is used in natural language problems like OCR or machine translation.

Maybe google "machine translation segmentation" or OCR for references to papers on the topic, but methods for doing this are really successful (especially on english).

It took me awhile to get, but my aha moment was understanding its relationship to regular PaaS's (eg: heroku).

What more could you do if you didn't need to deploy your entire stack with each little change you make? If you could only deploy your frontend (in different versions simultaneously) pointing to a backend that's always up, it enables much faster iteration: testing, bugfixing, demoing things that otherwise you might not deploy different versions for because it's too painful.

For example, you could A/B test meaningful portions of frontend code in a more scalable way than, say, Optimizely. Or you might point a user that's having problems to a specific build (which you can deploy instantly) and see if it fixes their problem _on their live account_. Or you can hack together a new feature that works _on a live account_ for a presentation, and push it without worrying about mucking up or interrupting production traffic. Or turn off minification instantly and deploy that.

Plus, you don't have to worry about refining internal frontend build scripts, it's plug and play.

Just out of curiosity - do you mean a longer track record for us? You don't need to consolidate to use the site, but we've been getting positive reviews from some industry leaders since we launched after YC in S2010: https://www.readyforzero.com/praise-and-press

If you have any specific questions, email contact@readyforzero.com.

:) hehe - the video is a little silly, but it was a pretty fun show to do.

This is a good point. It's worth noting that much of that is boilerplate for bootstrapping and datastructures, and that further code additionals would result in a more reasonable size increase. This thread mentions some of the reasons: https://groups.google.com/forum/?fromgroups=#!topic/clojure/...

Sennheisers were standard issue at Google for a long time, I wonder if we could normalize by the number of Google hackers.

Any good examples of some JS that's cool to run?

I can assure you that there exist some scenarios in which it will take more than 10 minutes. However I can also assure you that there exist some scenarios in which it will take less than 10 minutes. Thanks for trying it out!

Thanks!

Thanks for pointing that out. Yes, we could have checked every single corner case, but our job is to build financial software, not programming challenges, so we did it quick and dirty :)

:) I meant linear, but I'd review a proof :)

Not an SEO experiment - trying to hire smrt people!

The characters should be interpreted as ASCII.

Can you do it in linear time in the length of the input?

Very good points. The computer screen wasn't that large so I tried to keep the descriptions short. :)

you're right - we should explain that it's balanced.

If anyone wants the hard version, let me know (you can upvote this comment)

I think that's the spirit of the post, that building persuasive systems is moving/has moved into the real world in a principled way.

Disclaimer: I work with the author, and it's suprised me how much of a 'science' it really is (hypotheses can be tested, and the results can be measured and improved)

My intent there was not to say that they are key in python, but rather in more functional languages.

Yet another reason I worry that Python 3 may fracture the community...

would be cool if it supported mms.

I remember then that they had to put up a sign: "Don't be evil, please take 1"

you'll still need to send it. depending on the users you're supporting, and how, yodlee supports the long tail (some better, some worse), which would be difficult to support in a model like this.

Every time I've tried comparing cuckoo hashing vs traditional hash algorithms in practice, the time taken to compute the additional hash functions outweighs any gains in performance.

Counter-intuitively, I've also noticed in many cases that using binary search over sorted elements in contiguous memory is actually faster than using a hash table at all.

Has anyone else found this?

One interesting tidbit: the most popular support request we've received since launching is to make the service available in the US and Canada.

One natural reason that this situation has gotten worse is that the credit card market has become completely saturated (everyone who wants a card has one [almost]). This, combined with relatively easy credit recently has resulted in the card companies marketing to people that they would previously had stayed away from.

We've learned that people often become accustomed to credit at an early point in their lives (college), rely on it for awhile, then don't suffer any explicit repercussions, and so grow accustomed to it until the problem grows and it's too late.

I don't really think it's a 'conspiracy' either, since it's clear what the motivations for all of the involved parties are.

Would love to chat about this more, but I need to tend to the serverz :)

The most recent Survey of Consumer Payment Choice estimates that there are 178M card holders. The 2009 Survey of Consumer Finances (2) estimates that percentage of card holders of "bank-type" (eg: not Macy's) cards that revolve balances is just under 50%. With store cards, this number is right around 100M.

(1) http://www.bos.frb.org/economic/ppdp/2009/ppdp0910.htm (2) www.federalreserve.gov/pubs/bulletin/2009/pdf/scf09.pdf