Columbia MAFN Some Information about Columbia MAFN

Omar Felix

Member
I just finished my first two semesters in the MAFN program, and I'd like to share some information about the program.
I am an international student, thus my information may not be that suitable if you are an American citizen.

1. Several things need to be updated.

a. Thanks to the new director Lars Nielsen, now we have CPT in the summer and the third semester, which is of great importance for the international students doing summer internships and part time jobs in fall. Lars is a great director, and everyone can see his effort to make this program better and better.

b. If you choose to graduate after three semesters, in the third semester the tuition is about 10k, and you can still take up to 18 credits of courses.

c. During the three semesters, in total we can take 54 credits of courses. If your math or computer science background is not solid enough, you can choose undergraduate level courses in the math or cs department. The course registration is really flexible. I even took an advanced French language course cuz I had a double major in French at my undergraduate school. I also saw people taking courses from the MFE program or the business school, which is not easy but possible.

d. If you don't want to waste you 18 credits in each semester, you can choose 6 courses and take two of them with R credit, so that you can learn something and do not need to take the exam at the same time. You won't have any grades on courses of R credit.

2. About the courses.

In general the courses in the program are not as good as I thought. The curriculum is very badly constructed. For example, in the first semester, on the Stochastic Process course, we did not learn a lot of things about Stochastic Process, the course was mainly about stochastic calculus. Even now I am still not that familiar with things like Poisson Process or Markov Chain cuz they were not covered in the Stochastic Process course. Then, in the second semester, the Stochastic Methods course just started all over again from the beginning of the stochastic calculus theory, which was really a waste of time. It was actually easier than the Stochastic Process course in the first semester. The funny thing is, at the same time, the Nonlinear Option Pricing course in the second semester started with the Generalized Feynman Kac theorem, which was introduced a little bit in the last class of Stochastic Method. Therefore, I was totally in a mess: in the first week, the professor of Nonlinear Option Pricing used Feynman Kac Theorem, which I learned 15 weeks later. You can imagine how confused and struggling I was during the whole semester.

I can give you another example. When I was informed that MAFN would have its own C++ course, I was really exited because my programming background was really weak. Then, the fact turned out to be really cruel: the course was like another high-level math course. The professor just talked a little bit about the C++ syntax in each class, and then move on to solve really complicated Partial Integro-Differential Equations, which more than half of the students never learned before. He also used very advanced techniques like Fast Fourier Transforms to price options. I literally had no idea about what he was doing at that time, and the whole semester was like a nightmare. In the final exam I got less than 60 out of 100, I was very sad until I met an American girl who got less than 20. The good thing is, in the second semester, I learned from the Numeric Methods course about how to solve PDE numerically, the relationship between PDE and SDE, and the tricks to solve the inverse of a tridiagonal matrix, which were already regarded as "prerequisites" of the C++ course in the first semester. The sad thing is that even now I still have no idea about Fast Fourier Transform.

Also, some of the contents of the courses are way too theoretical. In the Stochastic Process course, it took me a long time to understand the measurability of a stochastic process, and then there will be a line on the notes saying "the measurability of a stochastic process does not have any applications in the finance industry", it's only for the completeness of the mathematical theory.

I can give you another example, in the final exam, given two stochastic processes, we were asked to distinguish the difference among "almost everywhere identical", "stochastically equivalent", and "indistinguishable", which also has no financial meanings at all. Such contents are not rare at all in the MAFN courses.

In short, most of the courses are really tough, practical and sometimes too theoretical. We can learn a lot of things. I got one of my internship offers because I did a project to determine the parameters of OU process by Calmen Filter and EM algorithm, which is exactly what the internship will be about. However, I really hope the professors could sit down together and have a conversation with each other, so that they will understand the whole structure of the curriculum and have more ideas about what should be included in their own courses. Of course, if you are already a Math or CS PhD, then this program will be like a paradise for you. The curriculum is just too awkward for students with weak background like me.

3. About the career service.
I know about 50 students in the program, and only 5 of them found the internships or the full time positions through the program. Most of us found jobs or summer internships by ourselves. However, you know, this is NYC, and Columbia's name is big enough. It was miserable to look for summer internships all by myself, but I finally got three offers after applying for more than 150 positions. So it's not as hard as I imagined. Also there are students who did not get any summer internships at all. I am not very familiar with everyone of them, but I guess the reason might be like following:
a. Wasn't going to look for jobs or internships at the first place.
b. English might not be good enough.
b. Skill set might not be solid enough.
c. Not familiar enough with the questions in the interview books.
d. Not lucky enough. I have to mention that I am fluent in French, and I always met French managers in the interviews. Seeing a Chinese student speaking fluently in French really amazed them, which also made the atmosphere of the conversation really comfortable and amiable. I can't guarantee the result will be the same if the managers were like Russian or Indian, since I do not speak Russian or Hindi.

Here is a list of the placement data of the 14 fall MAFN students. Please note that I only asked about 40 students. All of them are not US citizens, and there are about five students who found internships but I don't know which companies.

Company Position Location
Goldman Sachs, Strats Summer Intern, HK
XL Group, Inverstment Summer Intern, NYC
Chase, Quant Research Summer Intern, NYC
JP Morgan, Fixed Income Full Time, NYC
JP Morgan, Quant Research Summer Intern, NYC
JP Morgan, Market Risk Summer Intern, NYC
Credit Suisse, Technology Summer Intern, NYC
Bank of New York Mellon, Asset Management Summer Intern, NYC
SunGard Risk Consultants, Summer Data Analyst, NYC
Stevens Capital Management, Quant Research Summer Analyst, PA
AIG, Actuarial Summer Intenrn, CA
AIG, Quant Research Summer Intern, NYC
Credit Suisse, IBD Summer Intern, HK
HSBC, S & T Summer Intern, HK
Societe Generale, Technology Summer Intern, NYC
Citi, S&T Quant Research Summer Intern, NYC
Trexquant, Quant Research Summer Intern, CT
Trexquant, Quant Research Summer Intern, CT
Moody's, Financial Modeling, NYC
Moody's, Financial Modeling, NYC
Barclay's, Market Risk, NYC
Black Rock, Quantitative Research Summer Intern, SF
Black Rock, Quantitative Research Full Time, NYC
Bank of America, Quantitative Research Full Time, NYC
This is not an official version of the placement. Forgive me if I make some tiny mistakes about the positions' exact names.

Finally, I have a suggestion for international students. If you want to stay in America, you should literally graduate after three semesters. This is not only about the importance of summer internships. This is about how many times you can try H1B lottery. The application of H1B needs to be submitted in April, if you graduate in December, within 29-month OPT, you will have the opportunity to try the H1B lottery for three times. But if you graduate in May, then you'll only have two times to try. By the way, in order to have 29-month opt, try to find a job in an E Verifying company. AIG and Societe Generale are not E Verifying companies so you'll only have 12-month opt if you work for them.

I tried my best to expose both the positive and the negative parts of the program, and I hope this can help you if you are considering to choose this program. ;)

Upadate: i got a question like this: do you think Columbia's mafn deserves it number 7 rank? My answer is: absolutely. I don't think the programs ranked after MAFN are doing a better job. I have friends in other programs, and I also met some students from other programs during the interviews. As far as I can see, most of them are in a even worse situation. For example, I have friends from the NYU MathFin program and CMU still didnot find any paid internships. These two programs are good, but not as good as I imagined. Not to mention the programs like Cornell MFE, Chicago MathFin, or UT MathFin. Columbia's big name the location, and the director's effort kinda offset the short comings I mentioned. If you let me choose programs once again, i will only change my decision if I got offers from Columbia MFE, Priceton MathFin, or Berkeley MFE. In short, location and Ivy League's name does matter. PS: several students in 13 fall got associate offers from top invest banks, I know one student who declined the offers of Quant Associate position from JPM and MS, and finally joined in a hedge fund.
 
Last edited:

IntoDarkness

Well-Known Member
thx for sharing... ironically, i found the mfe program in the engineering school is pretty well structured and taught...
 

Omar Felix

Member
thx for sharing... ironically, i found the mfe program in the engineering school is pretty well structured and taught...
why is that ironic? The MFE program under IEOR is the best MFE program with no doubt. Even CMU or Baruch cannot compare with it.
 

anonymecon

New Member
This is an amazing post! Thanks for sharing such aboundant evidence and details.

Btw, personally I don't think it's a problem that some courses take quite a few advanced topics as their prerequisites.
It's never hard for us to learn sth. by ourselves, is it?

A question:
I heard that some students are able to find part-time internships in the second semester(the spring one) after enrollment? Why is it possible for international students holding F-1 Visa?

I'd be grateful if you could help with my confusion.
 

Christopher lee

Member
C++ Student
Hi Omar. Did the students who got these internships/jobs have prior job experience in the field? Also, what was your background undergrad? I was a math major who did well in diff eq, linear algebra, probability, etc, but did terribly on real analysis, abstract algebra, aka all the proof classes. Also, when did you start applying for internships? Finally, I read on columbia's website that we need to maintain a 3.0. How many people flunk out and how much does our gpa matter?
 

Omar Felix

Member
This is an amazing post! Thanks for sharing such aboundant evidence and details.

Btw, personally I don't think it's a problem that some courses take quite a few advanced topics as their prerequisites.
It's never hard for us to learn sth. by ourselves, is it?

A question:
I heard that some students are able to find part-time internships in the second semester(the spring one) after enrollment? Why is it possible for international students holding F-1 Visa?

I'd be grateful if you could help with my confusion.
It's unpaid.
 

Omar Felix

Member
Hi Omar. Did the students who got these internships/jobs have prior job experience in the field? Also, what was your background undergrad? I was a math major who did well in diff eq, linear algebra, probability, etc, but did terribly on real analysis, abstract algebra, aka all the proof classes. Also, when did you start applying for internships? Finally, I read on columbia's website that we need to maintain a 3.0. How many people flunk out and how much does our gpa matter?
All the students i mentioned did not have any full time experience.
I studied Finanial Engineering and French as double major.
real analysis is important to understand stochastic calculus.
it's curved. so it's easy to get gpa higher than 3.0.
 

Omar Felix

Member
well, i wouldn't go that far saying columbia mfe is the best
When I got my first internship offer, more than 70% of students in MAFN were still looking for summer internships.
At the same time, 80% of the Columbia MFE students already got their offers, while in CMU, CornelL MFE,and NYU math fin the number was 60%, 25%, and 40%. I got these figures by asking the students from the programs directly. Maybe it's biased.
 
Last edited:

pingu

Well-Known Member
When I got my first internship offer, more than 70% of students in MAFN were still looking for summer internships.
At the same time, 80% of the Columbia MFE students already got their offers, while in CMU, CornelL MFE,and NYU math fin the number was 60%, 25%, and 40%. I got these figures by asking the students from the programs directly. Maybe it's biased.
What did you hear from Baruch students?
 

MartingaleOpt

New Member
As another Fall MAFN grad, currently employed with JPMorgan Quant Research full-time, I just wanted to clarify a couple of things regarding the courses to provide a bit of detail to Omar's post.

The initial Stochastic Processes course really was quite poor in terms of teaching in my opinion. The implied aim of the course seemed to be a blend between theory and practice, but it ended up failing to be both. It was not practical as we learned essentially nothing about stochastic processes other than Geometric Brownian Motion. At the same time, theorem *statements* were given precisely, but since no proofs were involved, even for those who appreciate mathematical rigour, the course left much to be desired. On the contrary, this seems to be simply a function of the instructor (Lars Nielsen). In the winter semester, the course was taught by Jose Blanchet and was much more applicable to finance. The course covered Markov Chains/Processes, Feynman-Kac, Jump Processes, etc.

Non-Linear Option Pricing is by far the best course in the program. It is *very* difficult (and I say this having taken the STAT PhD level Probability I/II , Stat Inference I/II as electives). But at the same time it is very rewarding and the techniques covered in this class are very hard to come by in any program, including PhD programs in mathematical finance. Highly recommend taking it if it is offered again.

Also, this may not be the norm, but I was very successful getting interviews through the resume book service offered by the program. In total I interviewed for approximately 30 different jobs in the quant sector, about 10 of which came through the resume book service through no effort of my own - this is how I got the interview to my current position.
 
When I got my first internship offer, more than 70% of students in MAFN were still looking for summer internships.
At the same time, 80% of the Columbia MFE students already got their offers, while in CMU, CornelL MFE,and NYU math fin the number was 60%, 25%, and 40%. I got these figures by asking the students from the programs directly. Maybe it's biased.
For CMU, it's around 95% now and I know at least 9 people got the JPM S&T offer in HK but rejected the offers for better positions. This year's internship report is even sicker than last year's if you look at the positions. I asked a former Columbia MFE graduate, a good friend of mine, in 2014 and showed him CMU's employment report when I was deciding where to go. He said the Columbia website was not accurate because a lot of MFE students decided to stay for 3 semesters. After reading the report, he said I should go to CMU instead of Columbia. lol........Anyway, when I was deciding between these programs I joined some social network groups then I realized it's the exact same people in these different groups for Columbia, CMU, UCB, NYU and etc....A log of people got offers from these schools simultaneously so I doubt there will be a big difference......

As for brand name, we're financial engineers so perform or perish, no body cares too much where you come from like in IB/consulting......
 

Omar Felix

Member
For CMU, it's around 95% now and I know at least 9 people got the JPM S&T offer in HK but rejected the offers for better positions. This year's internship report is even sicker than last year's if you look at the positions. I asked a former Columbia MFE graduate, a good friend of mine, in 2014 and showed him CMU's employment report when I was deciding where to go. He said the Columbia website was not accurate because a lot of MFE students decided to stay for 3 semesters. After reading the report, he said I should go to CMU instead of Columbia. lol........Anyway, when I was deciding between these programs I joined some social network groups then I realized it's the exact same people in these different groups for Columbia, CMU, UCB, NYU and etc....A log of people got offers from these schools simultaneously so I doubt there will be a big difference......

As for brand name, we're financial engineers so perform or perish, no body cares too much where you come from like in IB/consulting......
You can certainly choose CMU if you don't care about the big name of Ivy League. Everyone may have a unique preference.
 
You can certainly choose CMU if you don't care about the big name of Ivy League. Everyone may have a unique preference.
What exactly "big name" are you talking about? I understand people go to Columbia for their campus environment/legacy/professors and so on, but if a financial engineer would rely on the name of their school to get jobs then he should probably look for another career. Ivy League undergrad certainly deserve its reputation (I went to a non-ivy but equally prestigious college so may be I just don't care about brand name), but not so true for their masters. This is especially true for Columbia because it has soooooooo many master programs. To be honest with you I know many people who didn't go to Columbia exactly because it's Columbia. The fact that they have such a low admission bar for so many mediocre master programs (eg. the Chinese dominated Statistics MA) has already severely damaged their brand name, especially in Mainland China. This is even affecting its top tier programs like MFE. If Columbia doesn't stop doing so, its reputation would be ruined further.

After all, we're financial engineers, so perform or perish. There's no other way around. Your future track record as a PM, trading book PnL as trader/strat, and research projects as an analyst are the only things that matter.
 

IntoDarkness

Well-Known Member
well, brand name definitely helps get into the door. to many ppl, that's all it matters. after that, future success depends on so many things... luck arguably plays a much bigger role than one's skills...
 

Reyn S

Member
Well, it's more than one month, I still got no message about my waitinglist.
 
Top