What is an entry-level trader's annual salary supposed to be?

[IlyaKEightSix]

1) Ahhh...Radix sort. Just wiki'd that one up. Never saw it in class before.

2) Merge vs. Quick: depends on the data structure. Because merge sort is a stable sort and quick isn't, quick would go better on arrays of numbers, while merge would do better on more complex data structures, such as linked lists.

3) For one linked list in O(n): go down the list, creating a double link at each node, then delete the original.

4) If the two linked lists merge at the last node, then they're one big linked list, and thus can be manipulated as such.

5) Subset sum: ah, the Horowitz and Sani solution. Split the elements into two arrays, sort the arrays, then with one, start from the largest value, and from the second, start from the smallest. Keep adding them up simultaneously until you reach above the desired target, then subtract from the first while adding from the second. This runs in O(n sqrt(2^n)). I would wonder if it can be improved by sorting all the numbers initially, and then dividing them into a smaller and larger array at the median.

6) public vs. private pedals: hmmm...I know that a private method can only be accessed by instances of its own class. So an instance of mountainBike, I'd think, would have no problem accessing its pedal() method. Furthermore, since bike's pedal() is public, it can be accessed by any bike class. That said, I think that if you tried to do it in reverse, and made Bike.pedal() private and mountainBike.pedal() public, that mountainBike.pedal() wouldn't be able to access the original pedal() method and the compiler would give you an error. It's been a while since I remember the exact wording of the error, but basically that you can't access a private method from outside its class.

 

[GoIllini]

1) Ahhh...Radix sort. Just wiki'd that one up. Never saw it in class before.

Good. Be advised wikipedia will not be available in most interviews.

2) Merge vs. Quick: depends on the data structure. Because merge sort is a stable sort and quick isn't, quick would go better on arrays of numbers, while merge would do better on more complex data structures, such as linked lists.

Any comments on memory usage? Worst-case sorting times? What happens if the numbers may be partially or fully sorted?

3) For one linked list in O(n): go down the list, creating a double link at each node, then delete the original.

Good. Now let's assume the link node class doesn't have a variable for a second link. How do you do it? Be ready to write out the code for this.

4) If the two linked lists merge at the last node, then they're one big linked list, and thus can be manipulated as such.

Let's say it's a single-direction linked list rather than a two-direction linked-list.

5) Subset sum: ah, the Horowitz and Sani solution. Split the elements into two arrays, sort the arrays, then with one, start from the largest value, and from the second, start from the smallest. Keep adding them up simultaneously until you reach above the desired target, then subtract from the first while adding from the second. This runs in O(n sqrt(2^n)). I would wonder if it can be improved by sorting all the numbers initially, and then dividing them into a smaller and larger array at the median.

1.) You Wikipedia'd that, too. I've never seen an algorithms prof call it the Horowitz and Sahni solution. Or apparently, the profs at Illinois hate Cornell.
2.) That was one of FIVE solutions they proposed for avg-case O(2^.5n time) (My thesis was a sixth solution- that's the only reason I know Horowitz and Sahni)
3.) You gave me a solution that doesn't really help you here- it's still not a P-time solution. HINT: is there another algorithm they propose that breaks the numbers down into discretable components?
4.) You didn't answer my question. Is there a polynomial time solution for solving the problem on natural numbers (positive integers)?

6) public vs. private pedals: hmmm...I know that a private method can only be accessed by instances of its own class. So an instance of mountainBike, I'd think, would have no problem accessing its pedal() method. Furthermore, since bike's pedal() is public, it can be accessed by any bike class. That said, I think that if you tried to do it in reverse, and made Bike.pedal() private and mountainBike.pedal() public, that mountainBike.pedal() wouldn't be able to access the original pedal() method and the compiler would give you an error. It's been a while since I remember the exact wording of the error, but basically that you can't access a private method from outside its class.

Yes. I'm not asking about who can call a private method. Can you override a private method with a public method? Can you override a public method with a private method?

 

[vkaul]

Are these programming expectation for all quant jobs in wall street? It is very CS oriented,right?

 

[IlyaKEightSix]

Are these programming expectation for all quant jobs in wall street? It is very CS oriented,right?

Not sure, but might as well try and cram some of it down so I don't look like a fool when it comes to those interviews. A lot of the questions GoIllini is asking right now are things I'm seeing for the first time.

Reminds me of the fact that I met a director from Google NYC who is actually in his spare time teaching middle schoolers how to program.

If only I found out how useful it was back then!

 

[GoIllini]

Are these programming expectation for all quant jobs in wall street? It is very CS oriented,right?

No, but if you're getting hired as a programmer, there's an expectation of a stronger skill-set with technology and algorithms to make up for an expectation of a weaker skill-set with some of the quant stuff. It's impossible for most banks to find enough people who are incredibly strong at both CS and financial engineering, but if you're not going for a quant or trading position, you'd better know your CS stuff.

Not sure, but might as well try and cram some of it down so I don't look like a fool when it comes to those interviews. A lot of the questions GoIllini is asking right now are things I'm seeing for the first time.

Reminds me of the fact that I met a director from Google NYC who is actually in his spare time teaching middle schoolers how to program.

If only I found out how useful it was back then!

You're actually doing OK; I'm just throwing the gauntlet at you for a traditional financial programming/ tech company interview. Assuming you're not using Google too much, I think you can land a solid job as a programmer in IT at an F500 company if you get a few interviews and can act non-typical-Ilya-on-the-boards in them.

Focus on the easy stuff- data structures and sorting- if possible, have a programmer friend drill you on this stuff, so you're ready. That's 30-50% of a technology interview, and if you can manage that stuff well, you won't look that bad. And PLEASE PLEASE PLEASE, whatever you do, just because you "crammed it down" a few weeks before doesn't mean you should say yes if they ask you if you "know algorithms."

 

[IlyaKEightSix]

Oh, heh, one update...I did get an email from someone at Google asking for my transcript screenshots, and that he was so happy I had such an interest. Not an offer or an interview even by any stretch of the imagination, but...

WOOT! I'm hoping my 4.0 gets me to the interview, because if I can go down to Mountain View or even up to NYC over spring break... ...

So far, there's one position I might be able to have a shot at (decision support analyst), but if (and that's a *huge* if) I can get in (go go go 12 credits of R programming!), 3-4 years down the line, Google has their own vaunted Quantitative Analyst positions.

In the case of Google, I have two cards to play:

1) Russian immigrant at young age who likes numbers (I think Google knows about those cases quite well)

2) 12 credits of R programming!

*Fingers crossed*

 

[atreides]

good luck buddy

 

[GoIllini]

Oh, heh, one update...I did get an email from someone at Google asking for my transcript screenshots, and that he was so happy I had such an interest. Not an offer or an interview even by any stretch of the imagination, but...

WOOT! I'm hoping my 4.0 gets me to the interview, because if I can go down to Mountain View or even up to NYC over spring break... ...

So far, there's one position I might be able to have a shot at (decision support analyst), but if (and that's a *huge* if) I can get in (go go go 12 credits of R programming!), 3-4 years down the line, Google has their own vaunted Quantitative Analyst positions.

In the case of Google, I have two cards to play:

1) Russian immigrant at young age who likes numbers (I think Google knows about those cases quite well)

2) 12 credits of R programming!

*Fingers crossed*

Only a few people can get hired into Google or any firm because they're the best developers out there, and these people usually have degrees that say PhD, MS CS, or MIT and IIT on them. You need to avoid competing against on that basis. In fact, from one state-schooler to another, I'd advise you to not count on winning a head-to-head competition against stats majors from Harvard and Princeton.

Figure out your niche. For me it was algorithms. For you, it's probably a specialized area in stats. Throughout the entire interview, you will need to downplay your background, but try to answer every question they throw at you, and then say that they really want to hire you because of your background in XYZ within stats. I was up against kids from MIT and CMU, but one firm decided it needed someone who specialized in algorithms more than it needed a general programming savant, so I got the offer.

Even if you're God's gift to the programming or stats world (probably about as unlikely as it is for me), that doesn't mean they're going to hire you. You're up against other people who are also the best thing to happen to computers since Babbage, it gets hard for some of these interviewers to compare candidates, and you're only one out of five or six. Figure out a way to work in a clear reason to hire you that plays to your strengths. I find myself mentioning algorithms a couple of times in every interview, and sometimes in the thank-you note when it doesn't seem overbearing to bring it up yet again.

Don't play the "founder was a Russian immigrant too" card. Unless you're interviewing with a Russian immigrant, nobody cares, and even if you are, it should remain unsaid. Focus on where your intellect REALLY comes into play- and play that card.

Finally, you need to relax. It looks like you're trying way too hard, and you just need to calm down. Consider taking some St. John's Wort or getting a short-term prescription for an anti-anxiety medication to help you stay calm in your interviews and just focus on what you bring to the table that few others can. (Us Dutch Calvinist folks are very experienced at dealing with anxiety.) Try not to worry about the MIT kids because you're not competing against them on your merits as a developer but instead because of the fact *cross fingers* that Google can use your specialty within that stats area you focused on.

Just be the aloof guy who probably spent too many long nights working on XYZ in stats and is totally chill, and you'll make for a very strong niche candidate- assuming that niche needs filling at Google.

 

[Joy Pathak]

Since the original post started off with entry level traders annual salary..I had some additional questions. I went through most of the thread except for some of the questions on programming going back and forth.

What can be expected for a starting salary for an entry level jobs for a student with no full-time experience out of a MSF program. MSF- with FE concentration? with an undergrad in eng.

Any speculations or ideas? I figured ...for example at an average prop firm... 50K+Signing bonus + year end bonus is a reasonable expectation for the first 2 years at least?

|Just wondering.

Other than that.. Great thread.

 

[GoIllini]

Since the original post started off with entry level traders annual salary..I had some additional questions. I went through most of the thread except for some of the questions on programming going back and forth.

What can be expected for a starting salary for an entry level jobs for a student with no full-time experience out of a MSF program. MSF- with FE concentration? with an undergrad in eng.

Any speculations or ideas? I figured ...for example at an average prop firm... 50K+Signing bonus + year end bonus is a reasonable expectation for the first 2 years at least?

|Just wondering.

Other than that.. Great thread.

Depends on the firm. Most of the investment banks will hire you as a second-year analyst/employee with an MS, so you'd be looking at a 65K-75K 1st year salary + signing bonus + performance bonus.

Harder to tell with trading firms. Some of the crazier ones will pay you $40K + % of net P&L your first year on the job.

 

[Joy Pathak]

Haha. I like the 40K+ % of P&L. I like the Laundry Service and breakfast/lunch benefit packages of the prop firms in Chicago.

I interviewed with Morgan Stanley for their Model Review Group. I wonder how much they give. Will find out in few days about their decisions.

 

[IlyaKEightSix]

Just sent my resume to Omnis.

 

[Dibbs]

Just sent my resume to Omnis.

To how many places did you apply ? You're always talking that you apply to different places. Are you getting any results back ?

 

[IlyaKEightSix]

To how many places did you apply ? You're always talking that you apply to different places. Are you getting any results back ?

To how many places? I don't keep track. If I see a position I can possibly be of use to, I'll throw my resume.

As for Omnis...nope. No dice. They're just not hiring ATM. When they have stuff to do, they'll hire people...so they're opportunistic employers, as the rep told me.

Sucks, I know.

However, I did get an email from comScore (interwebz marketing) and I'll be giving them a call. Yeah, life sucks. Ever since I read that "one word for today's grads: statistics" article that stated Hal Varian (chief economist @ Google) said in ten years, stats was the place to be, I jumped in.

One thing I forgot to consider was...

What do I do between now and ten years from now...

Economy's still in the dumps. Wall Street's still more or less closed, everyone's still penny-pinching, and in order to gain experience, you have to have some to begin with, unless you're willing to work for minimum wage sans benefits (until I can otherwise note), which is a no-go for me due to loans/costs of living.

 

[atreides]

Okay dude, this might sound cheesy, but at some point everyone started at ground zero. So if you need to get experience to get that job, there a few things you could do.

You can go incorporate an LLC and look into sites like Elance | Outsource to freelancers, professionals, experts, and consultants - Get work done on Elance. and Outsource to Freelancers, IT Companies, Programmers, Web Designers from India, Russia, USA, and more - oDesk for contract type work if you have a marketable skill set. You can low-bid for projects or offer some free services to build a client base. Look on craigslist for paid projects, there is a section for services.

Look into jobs.joelonsoftware.com , they advertise some quant like jobs.

 

[IlyaKEightSix]

The problem being that I don't have an out-of-the-box marketable skill set. I need training. The way I always thought about is that in school, I'd prove I was trainable, get some experience working, and then do whatever (start my own firm).

But if I can't get off the ground...

 

[Andy Nguyen]

I've moved Ilya's posts from "Don't be Normal" thread to this thread dedicated to his job hunting effort.

Please do not hijack other people thread.

 

[atreides]

I always thought the goal of an interview with a prospective employer is to show them you have some of the skill set required to hit the ground running, and not to convince them you can be trained!!

The problem being that I don't have an out-of-the-box marketable skill set. I need training. The way I always thought about is that in school, I'd prove I was trainable, get some experience working, and then do whatever (start my own firm).

But if I can't get off the ground...

 

[Andy Nguyen]

Instead of beating this horse to death, you can improve your employability by proving you can do "stuff".

You have been saying your R skill is pretty good. Code up some project and put it here for people to use. If you are good, someone will hire you. Employers are looking to hire people who can do things, not to train them.

 

[DanM]

I think Ilya's never-ending quest for a quant job is worthy of its very own reality show.