How Twitch Learned to Make Better Predictions About Everything (2017)

Re:How Twitch Learned to Make Better Predictions About Every... by jnordwick on Tuesday March 20, 2018 @10:34AM (#16628891)

I once had am interview at an HFT firm that had a very long quiz. It was too long to finish completely and accurately in the time given (about twice as long - 45 minutes for about 20 questions plus 20 follow-ups), but you were given instructions to go as quickly as possible and finish as much as you can.

Every other question, a follow up to the substantive quotation, asked you to evaluate how sure you were in the previous answer.

So a question might be a little theory or short answer or maybe asked to write code to do a simple bloom filter. The next question asked how sure you were it had no bugs or would work if all values were 4 character strings.

Probably the second toughest interview I ever went on. But I found the idea fascinating. The idea was for the quiz to test your ability to work under pressure and be able to evaluate risks in your code.

[ Parent | on HN ]

Re:How Twitch Learned to Make Better Predictions About Every... by ghostbrainalpha on Tuesday March 20, 2018 @12:29PM (#16630061)

That sounds horrible and fantastic at the same time.

But what was the toughest interview if that wasn't it?

[ Parent | on HN ]

Re:How Twitch Learned to Make Better Predictions About Every... by jnordwick on Tuesday March 20, 2018 @12:45PM (#16630243)

Tie between a trading firm and a bulge-bracket bank. The trading firm was two days onsite. The first day was a lot of tough programming questions and the second day was tele-conferencing with other offices including regulatory and trading questions. (They did fly me out first-class and put me up in an amazing king suite at the Trump with full kitchen and living room).

The bank's questions were just brutally difficult. From low-level C++ and Java internals to algorithmic ones where you are at best hoping for a good approximation. And one very good series of questions on modeling a game where the answer involved using a stochastic matrix or monte carlo simulations.

[ Parent | on HN ]

Re:How Twitch Learned to Make Better Predictions About Every... by chasedehan on Tuesday March 20, 2018 @02:02PM (#16630968)

I also interviewed at some bulge-brackets - hands down the most difficult interviews I ever had (for quant roles). They kept hammering on really complex math/stats/programming questions. Once you would get one "right", they would hit you with something more difficult. I have read that their premise was to see how you could handle pressure and questions you didn't know the answer to. In one follow up I was asked if I remembered a previous question and if I had looked it up.

[ Parent | on HN ]