Forums  > Careers  > Quant Interview Study Guide  
     
Page 1 of 1
Display using:  

Jurassic


Total Posts: 379
Joined: Mar 2018
 
Posted: 2019-12-24 16:20
https://careers.google.com/how-we-hire/interview/#interviews-for-software-engineering-and-technical-roles

Google and Amazon both specify the topics to learn to their technical interviews. Also Steve Yegge also summarises this here. http://steve-yegge.blogspot.com/2008/03/get-that-job-at-google.html

Is there an equivalent for desk quant jobs anywhere? I find Joshi's Quant Interviews and Answers to be not very useful



nikol


Total Posts: 1234
Joined: Jun 2005
 
Posted: 2019-12-24 22:42
It looks like you have pretty wide range .. of interests.

Be careful, Google and Amazon hire with help of AI.
Amazon got even further - today, AI can fire as well...

Jurassic


Total Posts: 379
Joined: Mar 2018
 
Posted: 2019-12-25 21:40
@nikol side question - how many interests is too many?

nikol


Total Posts: 1234
Joined: Jun 2005
 
Posted: 2019-12-25 22:28
It depends on the number of talents the person has.

Modern definition of talent: young (and cheap). Those ads from recruitment companies: "we are searching for new talents" mean exactly these two features.

R.G.


Total Posts: 6
Joined: Jun 2017
 
Posted: 2020-11-13 00:46
"I find Joshi's Quant Interviews and Answers to be not very useful"
Can you elaborate?

I find Joshi a bit boring but still giving generally the right flavor of questions.
Zhou's "practical guide to quantitative finance" interviews is the best book I know of.
Scraping glassdoor of best companies worked for me okay a few years back, not sure how much of real questions is left there at this point.
Crack's "heard on the street" is a classic, feel somewhat obsolete but not completely useless.
Google "Interview Primer for Quantitative Finance", some guy went thru a bunch of interviews and posted a decent collection of questions on github (from writing feel he's not too sharp and some solutions are wrong and/or not very good - but questions are good).

Jurassic


Total Posts: 379
Joined: Mar 2018
 
Posted: 2020-11-13 11:07
@R.G.
""I find Joshi's Quant Interviews and Answers to be not very useful"
Can you elaborate?"

I just get questions from topics in the elements of statistical learning, various advanced statistics topics, and computer science topics. None of which are covered in there really.

"Google "Interview Primer for Quantitative Finance", some guy went thru a bunch of interviews and posted a decent collection of questions on github (from writing feel he's not too sharp and some solutions are wrong and/or not very good - but questions are good)."

Isnt this the guy with a phd in stats from oxford, who says 9/10 of quant interviews he doesnt get an offer. You would have thought a phd in stats should be enough to be a quant at most places.

gaj


Total Posts: 115
Joined: Apr 2018
 
Posted: 2020-11-13 11:38
> I just get questions from topics in the elements of statistical learning

You answered your own question. If you understand most of the stuff in Elements of Statistical Learning on an intuitive level, you're way ahead of most candidates.

Joshi's book is about derivative pricing. Very roughly speaking, derivatives pricing is for sell side quants, whereas stats and machine learning stuff is for buy side quants. You probably should figure out what role you're interviewing for.

sharpe_machine


Total Posts: 69
Joined: Feb 2018
 
Posted: 2020-11-13 12:20
>
Isnt this the guy with a phd in stats from oxford, who says 9/10 of quant interviews he doesnt get an offer. You would have thought a phd in stats should be enough to be a quant at most places.


Enough is enough, but top places are increasingly competitive to get into. So they have a 100s of qualified applicants each year and select only a very top ones. So, theoretically, if you are a top-20 in a cohort of quant-wannabees you would end up in a top-20 firm which might be orders of magnitude worse than FAANG.

R.G.


Total Posts: 6
Joined: Jun 2017
 
Posted: 2020-11-13 15:29
>I just get questions from topics in the elements of statistical learning, various advanced statistics topics, and computer science topics. None of which are covered in there really.

can you elaborate? (which funds, etc)
to be fair, some acquaintance with coding/algo questions on top of standard quant books is needed for a decent number of positions, I agree. as is really knowing how linear regression works, and maybe some pandas and basic practical data skills to be ready for that "data task".
beyond that, what you describe doesn't match my experience from either 1 and 3 years ago, and I interviewed pretty widely (mb ~100).
or my own interviewing people recently;)

I guess the biggest hidden variable might be the kinds of positions you're applying for/kinds of positions you're slotted into given your resume.
Certainly reading job descriptions could help here: there are 100500 different roles that can get labeled as "quant", from nlp researcher to c++ latency sensitive distributed systems dev to pdes and numerical methods whiz to excel monkey.

To be concrete, over the years I've interviewed with most sizable prop shops, many big banks, a decent number of multistrat hfs as well as some random hfs, some more hft firms (given my math phd, positions pry skewed derivatives related rather than pure stat signal search).
I think Zhou's book is not a bad approximation to the first PC of the questions I've been asked, even if it doesn't explain the majority of the variance.
I have some passing ESL knowledge, don't think was ever asked something more complicated than explaining L1 vs L2 regularization effects difference or smth else at the level no more than Ng's ML course, and even that felt off-hand. But regression questions are common (no "advanced statistics" though), as well as data tasks.

That said, I was never cool enough for mid-freq stat arb, and yeah from what I've heard they are more stats and ml heavy, given the nature of the business.



Jurassic


Total Posts: 379
Joined: Mar 2018
 
Posted: 2020-11-13 17:16
@R.G.

for example
i) l1 vs l2 regularisation like you
ii) kernel trick
iii) difference between continuous input and discrete input decision trees
iv) basic robust statistics
v) medium-hard leetcodes

a number of these were on the sell side as well

bandi_np


Total Posts: 7
Joined: Nov 2020
 
Posted: 2020-11-16 18:13
In general, you can find alternative solutions to the usual puzzles on math.stackexchange.com if the answers you find in the books or Glassdoor are not satisfactory.
Previous Thread :: Next Thread 
Page 1 of 1