I'm a senior buy side quant researcher. AMA

  • Thread starter Thread starter Igna
  • Start date Start date
Hi all,

As the title says, as prospective students should be making their decisions, I would like to use this platform to interact with students who might be interested in the career path.

Who Am I: I’m a senior quantitative researcher working in systematic equities. I’m what some might call a “full-stack” quant leading a team on the entire pipeline from data exploration to generating the trades that we want to do (not the actual execution though). I have been lurking on Quantnet for a few years. I did not pursue an MFE, but I did a related master’s degree and my choice for the school was informed by the rankings here. For anonymity reasons, this is not my main account.

About My Company: we are a billion$+ quantitative hedge fund dealing with all asset classes at various horizons. Our incoming researcher pool are heavily drawn from previous year’s interns.

Why Am I doing this (edited): thing can be a bit opaque from the other side, and I've enjoyed interacting with members both in this thread and in DM's. Would like to open a forum in an anonymous manner to give some perspective.

Ground rule: as long as there’s no question that can be used to identify me or my company, I’m ok to answer. I'm not going to endorse any particular schools/programs. Also, everything that I say is of my own personal opinion and there might be people with same title/responsibilities as me that have differing views. At the end of the day, I hope that what I'm saying is not totally idiosyncratic.

I'm on US Eastern time, and have a full time job so please understand if I don't answer immediately.

Pre-edit:
Why Am I doing this: I’ve been involved in our recruitment process of the last few years at all levels (undergrad all the way to experienced). As far as the process goes at our company, the chips are highly stacked against MFE students (because of how early the process starts). I personally would like to see more well prepared MFE graduates in our ranks, and I’d like to use this platform to both help students who are interested in that career path and collect some data points for my information.
 
Last edited by a moderator:
Thank you very much for this opportunity below is my background and questions.

Background

I am in Sixth form (Y12) in the UK . (So Grade 11 US). I study Mathematics, Further Mathematics, Physics and Computer Science

I recently learnt about this quant field and it interested me because I have always like math's, I started programming a few years ago really enjoy it and I recently started learning to trade which has exposed me to the financial markets and basic economics which I find interesting. So now I'm thinking a job in the quant sector would be amazing.

It is around the time now in the UK where we are being told to think about University and our careers so I would like some advice on what I should study.

I am also interested in electronic engineering in like robotics and devices

I have found different courses which I would be happy to study but none of them really have all of my interests because I would like to keep my options open.

These are links to some of the courses: electronic and electrical engineering, computing with finance and management, electronic and information engineering.

Questions

1. Is it best to study a degree with 1 job sector or career in mind or is me trying to keep my options open a good idea?
2. Can I get into the quant sector by doing engineering degrees like I have linked above and then doing a financial engineering masters?
3. What are the best courses I can study to get into the quant sector?

Thank you! :)

Wow, it's great to be prepared early. Since you are planning for it now before university, I would say plan your coursework so that you can get gather the necessary information to make a decision soonish over the course of your university life. I would also not advise not thinking about your course as the means to just getting a particular job or sector, because you might change your mind during the 4 years (or even after your university life). We have successful colleagues who have come from various qualitative backgrounds including engineering. My personal advice is to choose something that resonates the most with you, and spend your electives exploring other areas of potential interest.

If you want to maximize your chance of getting into a firm such as ours, I would recommend planning for an internship as a quant (researcher) in your last summer at the very least. We have an incoming intern this year who worked on robotics research the previous summer between his 2nd and 3rd years, and wants to get a feel for quant research. This is probably a smart thing to do since we mainly hire the intern pools and if he doesn't get the return offer or doesn't like us he still has the optional to do engineering.

By the time you are applying to the internship, you should've taken at least probability theory, statistics, linear algebra, multivariate calculus, and some programming. The basics aren't too much to ask for, but you need to make sure those foundations are solid. The most successful candidates also usually have specialized in a particular subfield (math/stats/cs/eng). That is usually demonstrated through graduate level coursework or some undergraduate research. This is a list of example classes that piqued my interest this past recruitment season, provided just to demonstrate the broadness of background, and not as actual recommendations: real/complex analysis, Fourier analysis, convex optimization, quantum computing, special relativity, nonlinear optics, reinforcement learning, Bayesian statistics, causal inference, computational complexity theory, computational chemistry, game theory, asset pricing theory, control theory, computer architecture, randomness and computation, etc.

Recruitment cycles start fairly early for us and other top tier competitors, which is early fall of junior year so it's definitely useful to plan ahead.
 
I'm aspiring to become a quant researcher in a buy side firm. I am an international student pursuing masters in computer science at george mason university. Right now I have one year left for graduation I have good understanding of machine learning, natural language processing and mathematics. Can you Please guide me how to get interviews for top tier firms and I have been googling a lot and couldn’t find a proper resource which can help me out. I recently came across the "wallstreetquant" bootcamp designed by the quants from wall street. The bootcamp is for three months and the price is around 4000$. I don’t understand if that is worth the amount. I am not able to figure out what skills and projects should I gain and work on, I want to master building the trading strategies, but don’t know a platform where I can learn them. I request you to please invest some time to guide me such that I get enough intuition on how everything related to quant works and I can start working on my profile.
I think a few of your skills are in vogue at buy side firms, namely ML and NLP. But the key questions are how are you setting yourself apart from rest of the field because these are also the courses that everyone else is taking, and how does it make you a better researcher. The first question could be answer potentially by whether you've taken more advanced courses than other people and know the theories in addition to applications, perhaps you've done some research or better yet have a publication. The second question is a reflection on your way of thinking and problem solving (i.e. start from a hypothesis, look for references, test it rigorously, question your conclusions).

Admittedly, getting an internship your last summer might present you with the greatest chance of joining as a fresh graduate. If you've missed that window of opportunity then I would put some thoughts and actions into how you can differentiate yourself based on the points in the first part of the answer and apply as a new graduate.

I don't know about the bootcamp you mentioned, and based on a quick look at its curriculum, I don't think it's really something that would necessary draw our attention. Sure it's nice to be better informed about what we do but that course is not going to make you a better potential researcher. Like I said in a previous post, fresh out of school you are not expected to be the strongest alpha researcher, and if a firm hires you with those expectations they are willing to give you the training.
 
Wow, it's great to be prepared early. Since you are planning for it now before university, I would say plan your coursework so that you can get gather the necessary information to make a decision soonish over the course of your university life. I would also not advise not thinking about your course as the means to just getting a particular job or sector, because you might change your mind during the 4 years (or even after your university life). We have successful colleagues who have come from various qualitative backgrounds including engineering. My personal advice is to choose something that resonates the most with you, and spend your electives exploring other areas of potential interest.

If you want to maximize your chance of getting into a firm such as ours, I would recommend planning for an internship as a quant (researcher) in your last summer at the very least. We have an incoming intern this year who worked on robotics research the previous summer between his 2nd and 3rd years, and wants to get a feel for quant research. This is probably a smart thing to do since we mainly hire the intern pools and if he doesn't get the return offer or doesn't like us he still has the optional to do engineering.

By the time you are applying to the internship, you should've taken at least probability theory, statistics, linear algebra, multivariate calculus, and some programming. The basics aren't too much to ask for, but you need to make sure those foundations are solid. The most successful candidates also usually have specialized in a particular subfield (math/stats/cs/eng). That is usually demonstrated through graduate level coursework or some undergraduate research. This is a list of example classes that piqued my interest this past recruitment season, provided just to demonstrate the broadness of background, and not as actual recommendations: real/complex analysis, Fourier analysis, convex optimization, quantum computing, special relativity, nonlinear optics, reinforcement learning, Bayesian statistics, causal inference, computational complexity theory, computational chemistry, game theory, asset pricing theory, control theory, computer architecture, randomness and computation, etc.

Recruitment cycles start fairly early for us and other top tier competitors, which is early fall of junior year so it's definitely useful to plan ahead.
Thank you so much for your insight. Seeing university as a way to pick up skills and knowledge instead of trying to get into a job sector has really cleared my head and made my decision easier.

I have also realised that engineering courses actually give you modules that include the stats, probability and linear algebra which will he needed for the quant field. Additionally the courses have electronic engineering (robotics and devices ) as that is the name of the course.

This keeps all my options open and allows me to gain experience through internships like you suggested, which can help me get a job like you said.

Again I am really grateful for using your time to answer my question 🙏
 
Hi @Igna, thank you so much for carving out the time to do this AMA!

Before asking my questions, I'd like to give some context. I'm a PhD student in CFD (computational fluid dynamics) which involves lots of solving PDEs numerically. My research is about completely redesigning and then parallelising CFD algorithms on the GPU to make them run really fast, so I'm more on the computational side rather than the mathematical side. I do a lot of coding using CUDA (if you know it), C++ and Python, using CUDA/C++ for high performance computing, and Python for data analysis and scripting. If it helps you to know more about my background, I've posted on here before.

I have two related questions as follows:
  1. Are my GPU computing skills valuable as a buy side researcher? I should clarify: excluding the sell side (because whenever I say PDEs, people point me to pricing in banks), and excluding quant developer (because whenever I say C++/CUDA people point me to development).
  2. Kind of related to question 1: are there quant research niches where my value add would be to interface between researchers and developers? My skill set boils down to having a good understanding of the underlying maths and translating it into fast code. However, I want to do research, not pure development.
 
Hi @Igna, thank you so much for carving out the time to do this AMA!

Before asking my questions, I'd like to give some context. I'm a PhD student in CFD (computational fluid dynamics) which involves lots of solving PDEs numerically. My research is about completely redesigning and then parallelising CFD algorithms on the GPU to make them run really fast, so I'm more on the computational side rather than the mathematical side. I do a lot of coding using CUDA (if you know it), C++ and Python, using CUDA/C++ for high performance computing, and Python for data analysis and scripting. If it helps you to know more about my background, I've posted on here before.

I have two related questions as follows:
  1. Are my GPU computing skills valuable as a buy side researcher? I should clarify: excluding the sell side (because whenever I say PDEs, people point me to pricing in banks), and excluding quant developer (because whenever I say C++/CUDA people point me to development).
  2. Kind of related to question 1: are there quant research niches where my value add would be to interface between researchers and developers? My skill set boils down to having a good understanding of the underlying maths and translating it into fast code. However, I want to do research, not pure development.
Hi alovya,

When it comes to hiring researchers we look at 3 angles: core competencies, comparative advantage and niches, with descending importance. Core competency are the usual math/cs/prob/stats areas. For mid frequency like what I do (minutes to days), C++/HPC falls under comparative advantage, while it falls a tiny bit more on the core competency side in higher frequency (under second). CUDA falls under a niche. I think there’s a bias barrier to overcome, and that is people who did research in a more computational discipline tend to not think about research the same way that we usual do in QR which is more like natural/experimental sciences (i.e. given a problem can you come up with hypotheses, given data can you test that hypothesis, and given results of your test can you propose actions).

Having said that, most PhD hires in researcher are taken on as generalists and as long as you can prove you have solid core competencies then you have a shot. You have to keep in mind that the pool of PhD applicants that we get is very diverse and there could be a lot of different skills being considered for each hiring season. In terms of GPU, our researchers do not need to go down to implementing in CUDA, can’t really comment on other firms but I’d imagine it’s probably the same. We use higher level abstractions like jax for the most part.

As for the second question, I think if you are interested in during quantitative research then you should focus on the relevant skillsets for research. It’s nice to have the ability to interface, but we can usually hire good enough developers who can bridge the gap and really want do development as well.

Based on the other thread you linked, I wanted to quickly comment on ML as well. We consider it part of the toolbox and if people have it on their resumes, they will get asked about it in fair amount of details. It will involve intuition regarding certain algorithms, pro and cons of certain assumptions and how it might translate to certain datasets with certain characteristics, very broad idea of learnability etc. (For example, an easier problem looks like: given a dataset drawn from a joint distribution P(X,y), with an MSE loss function and an algorithm that is infinitely flexible and guaranteed to converge to the global min, what is E[f(y|X)]? What if the loss function is absolute loss or asymmetric absolute loss? Can this E[f(y|X)] be achieved by algorithm XYZ?) The topics lean more towards theoretical ML in most graduate level programs and intuition of behaviors when you apply it to real problems. Then again, most people fall flat on some subtleties of OLS and multiple regression (which is assumed under stats core competency) before we even get anywhere close to nonlinear ML.

Hopefully this is helpful.

Igna
 
Hi Igna,

I really appreciate the thorough reply! Thank you for taking the time to read my posts in detail and give me tailored feedback on my situation as a computational PhD. I had a few minor follow up questions, if you don't mind.
I think there’s a bias barrier to overcome, and that is people who did research in a more computational discipline tend to not think about research the same way that we usual do in QR which is more like natural/experimental sciences (i.e. given a problem can you come up with hypotheses, given data can you test that hypothesis, and given results of your test can you propose actions).
I'm grateful to you for pointing this out. I can see how a computational discipline leads to a lack of hypothesis generation and testing skills which are desirable in QR as it is too "deterministic". This is because it boils down to: 1) devise an idealised model, 2) code up the model, 3) run simulations with the model and 4) extract results from the simulations. However, the results are usually deterministic not stochastic in nature, so the workflow of "generate hypothesis and design experiment, do experiment and gather data, interrogate data and confirm/reject hypothesis" does not show up in a computational discipline, whereas it is always there in the natural sciences.
Having said that, most PhD hires in researcher are taken on as generalists and as long as you can prove you have solid core competencies then you have a shot.
You mentioned maths/coding/prob/stats as core competencies, but I would assume that for QR, hypothesis generation and testing skills are also core competencies no? Or do you mean to say that as long I have core maths/coding/prob/stats, and show promise for building good hypothesis generation + testing skills, this gives me a shot?
As for the second question, I think if you are interested in during quantitative research then you should focus on the relevant skillsets for research.
Based on what's been said so far re: my computational background, I think we can agree that what I'm missing from my research skill set is my ability to generate and test hypotheses. Do you have any recommendations on how I could train this skill in the context of quant finance? For example, interesting data sets over which to apply statistical learning methods (like performing a linear regression over stock price volatility and volume to see if volatility is positively correlated with volume).
Based on the other thread you linked, I wanted to quickly comment on ML as well. We consider it part of the toolbox and if people have it on their resumes, they will get asked about it in fair amount of details. It will involve intuition regarding certain algorithms, pro and cons of certain assumptions and how it might translate to certain datasets with certain characteristics, very broad idea of learnability etc. (For example, an easier problem looks like: given a dataset drawn from a joint distribution P(X,y), with an MSE loss function and an algorithm that is infinitely flexible and guaranteed to converge to the global min, what is E[f(y|X)]? What if the loss function is absolute loss or asymmetric absolute loss? Can this E[f(y|X)] be achieved by algorithm XYZ?) The topics lean more towards theoretical ML in most graduate level programs and intuition of behaviors when you apply it to real problems. Then again, most people fall flat on some subtleties of OLS and multiple regression (which is assumed under stats core competency) before we even get anywhere close to nonlinear ML.
With my PhD timeline (submission deadlines etc), I strongly doubt I'll be able to learn non-linear ML in time before QR applications open; at best I'll be able to thoroughly understand linear regression. Would you say this, along with training my ability to generate and test hypotheses, would give me a shot a QR roles?
Hopefully this is helpful.
It really has been! Sorry for my numerous questions, and thanks again for sharing your insight.
 
Last edited:
Hi Igna,

Thank you for hosting this AMA. I am an incoming quant at a hedge fund. I have 2 questions

1. I have no navigation in the Quant Industry. I would be very appreciative of any advice to thrive in the field as a beginner. Such as how you become "senior", how to ask "good" question?
2. I graduated with a Bachelor in CS with a mediocre GPA (3.48) and I think it is a negative spot on my resume. At what point will this will not be considered when applying for jobs (especially when we cannot disclose a lot on our resume and during interviews)? I also get the chance to do a part-time online master in CS at a top school, should I take it to boost my future employability?

Thank you!
 
You mentioned maths/coding/prob/stats as core competencies, but I would assume that for QR, hypothesis generation and testing skills are also core competencies no? Or do you mean to say that as long I have core maths/coding/prob/stats, and show promise for building good hypothesis generation + testing skills, this gives me a shot?
In terms of generating hypothesis it’s really something more of gauging how you think, so probably critical thinking should be added to core competency but it’s some much more of a soft skill. Most of the hypothesis testing is covered in undergraduate level stats (unless someone’s a stats PhD then we’d expect them to say more).

We don’t usually frame the probelm explicitly as form some hypothesis and test it, but might want to tease it out of a question. I.e. say we give you a hypothetical situation/dataset and ask you to think of an interesting way to use it. You might hypothesize that if you split the data some way then it introduces a meaningful difference, and you can suggest something as simple as a two sample test to test that hypothesis. We’ll then delve further from there usually—like what distributional assumptions you are making, what if you violate some assumption, etc. Then once we established that the statistical procedure used to study the data is sound then we would like to hear how you think about possible actions to take based on it.

Based on what's been said so far re: my computational background, I think we can agree that what I'm missing from my research skill set is my ability to generate and test hypotheses. Do you have any recommendations on how I could train this skill in the context of quant finance? For example, interesting data sets over which to apply statistical learning methods (like performing a linear regression over stock price volatility and volume to see if volatility is positively correlated with volume).
Working with real data helps. I think reading A bit out of your own area and read/critique other people’s papers can help too. I had someone who came into one of the interviews telling me how they’ve read so many psychology papers and told me of all these ridiculous sounding claims. The punchline was that he really questioned some of them enough to think more deeply about the inferences that they could draw from the data samples and wanted to learn more about their experiment design philosophy and tests to convince himself of the validity of the claims.

With my PhD timeline (submission deadlines etc), I strongly doubt I'll be able to learn non-linear ML in time before QR applications open; at best I'll be able to thoroughly understand linear regression. Would you say this, along with training my ability to generate and test hypotheses, would give me a shot a QR roles?
If you thoroughly understand linear regression then I think you’re already better prepared than major of people coming from similar backgrounds.
 
Hi Igna,

Thank you for hosting this AMA. I am an incoming quant at a hedge fund. I have 2 questions

1. I have no navigation in the Quant Industry. I would be very appreciative of any advice to thrive in the field as a beginner. Such as how you become "senior", how to ask "good" question?
2. I graduated with a Bachelor in CS with a mediocre GPA (3.48) and I think it is a negative spot on my resume. At what point will this will not be considered when applying for jobs (especially when we cannot disclose a lot on our resume and during interviews)? I also get the chance to do a part-time online master in CS at ta top school, should I take it to boost my future employability?

Thank you!
1. seniority means different things at different places. For example, in addition to just doing research, I have responsibilities in 1) portfolio management 2) overseeing the research agenda and junior researchers 3) involvement in a broader range of business areas (e.g. strategic tech development, etc). As you might have guessed from that we are not a multi manager setup, and it’s very different if you are just part of a small pod. Seniority usually comes from gaining trust from having completed your core responsibilities well, and the tendency to grow into more strategic responsibilities. As for good questions, I think early on you show focus on how to do things in the most correct way (do your day to day work well) and why things work the way they are (bigger picture). One advice I had early in my career is to not stop at a single level of questions, but think of it as an onion and keep asking the whys until you get to a root cause that is more fundamental.

2. I think as soon as you’ve started your first job GPA doesn’t really matter, I don‘t even recommend putting it in the resume. I know application forms ask for that information for experienced hires, but for me at least the relative importance decays very quickly after undergrad. So say if you are leaving your current job after 1 year, I might take that and the GPA together to make some assumptions about you, but if you’ve been there 3+ years I wouldn’t even look at your GPA. At that point your interview questions are also going to change from a set of canned questions to more regarding your work. Future education should always be done with a bigger goal (e.g. gain additional depth in a subject area) in mind, and I wouldn’t conflate that with masking your undergrad GPA.
 
In terms of generating hypothesis it’s really something more of gauging how you think, so probably critical thinking should be added to core competency but it’s some much more of a soft skill. Most of the hypothesis testing is covered in undergraduate level stats (unless someone’s a stats PhD then we’d expect them to say more).

We don’t usually frame the probelm explicitly as form some hypothesis and test it, but might want to tease it out of a question. I.e. say we give you a hypothetical situation/dataset and ask you to think of an interesting way to use it. You might hypothesize that if you split the data some way then it introduces a meaningful difference, and you can suggest something as simple as a two sample test to test that hypothesis. We’ll then delve further from there usually—like what distributional assumptions you are making, what if you violate some assumption, etc. Then once we established that the statistical procedure used to study the data is sound then we would like to hear how you think about possible actions to take based on it.


Working with real data helps. I think reading A bit out of your own area and read/critique other people’s papers can help too. I had someone who came into one of the interviews telling me how they’ve read so many psychology papers and told me of all these ridiculous sounding claims. The punchline was that he really questioned some of them enough to think more deeply about the inferences that they could draw from the data samples and wanted to learn more about their experiment design philosophy and tests to convince himself of the validity of the claims.


If you thoroughly understand linear regression then I think you’re already better prepared than major of people coming from similar backgrounds.
Great, thank you for elaborating and thank you again for your time. Hope you have a good evening :)
 
Last edited:
Hey @Igna thanks for spending your time to helping us out!
I will be joining an MFE program in Fall 23. As you suggested that October is quite early for a lot of MFE students to be prepared for quant interviews. So what are some things I could work on beforehand, to help me land an internship in October. I am an electronics engineering undergrad, so my programming and math skills are both decent, but still need some work. While my finance/econ would be my weaker aspect. Also if you could give me a walk through how the recruitment process for an intern works, would be great!
 
Hey @Igna thanks for spending your time to helping us out!
I will be joining an MFE program in Fall 23. As you suggested that October is quite early for a lot of MFE students to be prepared for quant interviews. So what are some things I could work on beforehand, to help me land an internship in October. I am an electronics engineering undergrad, so my programming and math skills are both decent, but still need some work. While my finance/econ would be my weaker aspect. Also if you could give me a walk through how the recruitment process for an intern works, would be great!
I‘d recommend starting to get your resume ready during the summer. At that time, you should already know what courses your program offers and should put prospective courses on your resume as well. A recommendation there is to put yourself in the shoes of someone in your same program and same background, how can you sell yourself over this alternative self. I’ve seen people send in cover letters which can be useful, but I tend to prefer reading a quick summary and objective on top of your resume. These are all stylistic choices and is up to you, but you should make sure they are ready to go as soon as the online postings are made available. Also, there are some companies have some way of expressing interest in the internship which opens up well before the process starts so you should do some research on your prospective targets and put that in if it’s available.

Postings and interviews could start around the time the semester starts, so may sure you have prepared for the canned questions in math/cs/prob/stats—there are other threads here that provide some resources and pointers. As an engineering major, programming and math background is usually decent, but I’d definitely suggest familiarizing yourself with the kinds of questions that are asked at quant interviews. One particular area is somewhat new—machine learning—and of interest to us that none of the established quant prep books cover. We’ll probably only ask in depth questions about ML if your resume suggests that you‘ve had nontrivial experience with it. I’ve covered some pointers in previous responses in terms of what we are looking for.

In terms of the process at least for us, we sent people to target schools to do recruitment, so if you are at one of these schools (even if it’s an undergrad event), try to get some face time, and make an impression. Otherwise if you are just applying online, then it will first be screened by HR—this is usually not too demanding, and it’s why you should have a strong resume ready sooner rather than later to get past this round. We have someone from the research side then give you a technical interview to make sure you have the basics. This is usually pretty short and covers some fundamentals. Once you are past that you get an invite for a super day. It’ll usually last better part of a day, and cover the technical aspects in more details than the technical screening. You might as get some questions about your own experiences so make sure you are prepared there as well (say if you say you had a project in your freshman year, it’s fair game to expect in depth questions there and saying you don’t remember the details raises flags).

For all the interviews we are still doing most of them virtual. Hopefully this doesn’t come across as a surprise, but we realize it’s not as easy to go over technical question (say if you want to write out some formula and vet it with us). So it’s really worthwhile to practice communicating your thoughts clearly.
 
Last edited:
Shout out to @Igna. Thank you so much for taking your time to help us out. I learned a lot already reading your answers to other people. Thank you so much.

I’m sorry about the following long post. The next 4 paragraphs are my background. You can probably skip some if you just wanna see my questions.

I am a 6th PhD student studying physics at a non-target university. I also have a PhD minor in statistics, with courses including theory of statistics, machine learning, statistical computing, fundamentals of optimization. Math has always been my favorite subject. I really liked quantitative finance, and want to pivot my knowledge and skills from math and physics to a quantitative researcher.

My research area is condensed matter physics and involves discovering and characterizing novel quantum magnetic materials, that could be useful in both fundamental understanding of physics and application in technologies, such as quantum computing. That being said, my research is mostly experimental (crystal growth, material characterization (XRD, SEM, AFM, Raman), magnetization measurement and has almost nothing to do with quant. We also do some calculations that involves modeling and computations, like band structure and magnon interactions, but we do these by collaborating with other theory groups. I have decent publications, 10 articles so far but I am only afraid my experimental papers are not appealing to quant people. That is, my research topic itself has almost nothing to do with quant finance.

In terms of math and programming, I have pretty good understanding of probability theory, linear algebra and calculus. I have meager understanding and a little experience of implementing machine learning. I have basic programming skills in Python, R and MATLAB. I also know a little bit of C and C++ but I don’t use them in my work. Knowledge of C and C++ is getting rusty. I have never taken a data structure and algorithms course, although I have been exposed to some of those algorithms here and there.

For financial experience, I have traded stocks from 2018 and spent a lot of time on it in 2020. I learned a lot of manual momentum trading. I am good at Mark Miniverni’s momentum trading strategies. I also tried to implement momentum trading strategies in Python and learned a lot. I That being said, I don’t have any formal education in finance. I don’t know much about financial derivatives, pricing theory, or portfolio optimization theory.

I started to know about quant 4 years ago and started doing more research about the quant career from this year. Now, I am aspiring to be a quant, more specifically a quantitative researcher. I started preparing this year from late February and applied 19 positions, all hedge funds. Among them, I got 1 OA which I passed but tanked the interview after two rounds. I realized I wasn’t prepared enough. It’s now mid April so I have about 5 months until mid September before I apply for internship or full time positions. I’m now studying machine learning 4-5 hours a day, mostly working through ISLR and ESL. I want to design a pathway (study & internship plan) to break into the quant career.

I am still working full time in my lab so suppose 4 hours study a day. I have about 5 months, or 600-700 hours. My goals for the next 5 months:
  1. Enrich my resume (may be personal projects, research projects with a professor, trading competitions, internships in quant research, quant trading, software engineering, or data science) to get interviews.
  2. Have enough knowledge foundations to pass the interviews and get offers.
My questions:
  1. Self study plan: if I were to self-study for the next 5 months before I apply for buy side QR roles, what do you recommend me to study, in addition to prob&stat, linear algebra, calculus)? What are absolutely necessary for a quant researcher? What are optional? I have listed possible areas of study based on my research, for your reference.
    1. Optimization in asset management.
    2. Stochastic calculus for finance, Steve Shreve volume 1 and 2
    3. Derivatives markets by McDonald
    4. Data structure and algorithms
    5. Python programming
    6. C++
    7. Machine learning
    8. Mining of massive data sets
    9. Time series analysis
  2. Off-summer internships: now, I don’t feel I am ready for QR interviews and there are not many positions available. The chance of getting a summer internship is bleak. Or is it just my illusion and there are actually some available and I should apply? Are there many internship opportunities in the off-summer season?
  3. Other internships: If I can’t find a QR internship soon, what’s the next best internship option I should try? SWE or data science? Which one would better prepare me for future quant job applications?
  4. Projects: if I can’t find an internship in any of these areas, I would like to work on a project with a professor, which topics do you think would make me stand out?
  5. internship after graduation: is it okay to look for internships after graduation? Like I graduate in May in 2024 and do an internship during summer 2024? Or should I apply for 2024 full time positions?
  6. Comparative advantages and niche: since you mentioned before, candidates should have strong core competency, comparative advantages and niche. That’s really enlightening so THANK YOU so much. I am wondering what my comparative advantages and niche are / could be. My comparative advantage could be my statistical / math knowledge but I don’t have much research experience like some PhD have their theses on ML or statistics. My niche might be experimental physics? But I don’t know know how that is exactly related to quantitative finance and could differentiate me from other PhDs in this field. Could you gave me some advice?
I really appreciate your time and any inputs. I also appreciate if you can point me to the right resource, if you or someone else have answered some of my questions. Thank you again!
 
Last edited:
Hi all,

As the title says, as prospective students should be making their decisions, I would like to use this platform to interact with students who might be interested in the career path.

Who Am I: I’m a senior quantitative researcher working in systematic equities. I’m what some might call a “full-stack” quant leading a team on the entire pipeline from data exploration to generating the trades that we want to do (not the actual execution though). I have been lurking on Quantnet for a few years. I did not pursue an MFE, but I did a related master’s degree and my choice for the school was informed by the rankings here. For anonymity reasons, this is not my main account.

About My Company: we are a billion$+ quantitative hedge fund dealing with all asset classes at various horizons. Our incoming researcher pool are heavily drawn from previous year’s interns.

Why Am I doing this (edited): thing can be a bit opaque from the other side, and I've enjoyed interacting with members both in this thread and in DM's. Would like to open a forum in an anonymous manner to give some perspective.

Ground rule: as long as there’s no question that can be used to identify me or my company, I’m ok to answer. I'm not going to endorse any particular schools/programs. Also, everything that I say is of my own personal opinion and there might be people with same title/responsibilities as me that have differing views. At the end of the day, I hope that what I'm saying is not totally idiosyncratic.

I'm on US Eastern time, and have a full time job so please understand if I don't answer immediately.

Pre-edit:
Why Am I doing this: I’ve been involved in our recruitment process of the last few years at all levels (undergrad all the way to experienced). As far as the process goes at our company, the chips are highly stacked against MFE students (because of how early the process starts). I personally would like to see more well prepared MFE graduates in our ranks, and I’d like to use this platform to both help students who are interested in that career path and collect some data points for my information.
View attachment 47121
Thank you so much for the post,

I hold a pure math degree from a CUNY (with 2.5 gpa), currently working toward a PhD (at a math program that ranks top 40s in US, considered a group I public by AMS) also with focus in pure math. But I decided I want to become a quant researcher when I graduate. So, I applied to a couple MFE programs and actually got admitted into some good program. But unlike PhD program are supported with funds, for me to attend any of the MFE program, I probably have to take out a big loan. So, I am contemplating whether or not I should attend a MFE or self-study. Also, I have not taken any CS courses.

So, I believe answers to the following questions will help me decide:

1) Will completing a MFE make up for the low undergrad gpa?

2) Based on your experience, what are some realistic starting salaries for quant researchers? Or what google says is accurate (~150k)?

3) Any advice that you could give?

Many Thanks!
 
(sorry for the delay, was traveling)
  1. Self study plan: if I were to self-study for the next 5 months before I apply for buy side QR roles, what do you recommend me to study, in addition to prob&stat, linear algebra, calculus)? What are absolutely necessary for a quant researcher? What are optional? I have listed possible areas of study based on my research, for your reference.
I would put some emphasis on implementation (data structure, algorithms, etc). Usually PhD are given a bit of benefit of doubt, but not being able to implement will be a blemish on the interviews, and will hobble you for the actual internship. At some point before you apply, you might want to learn a bit more about the target companies and the kind of questions that they ask. The breadth of knowledge, if you wanted to cover everything, is very vast. I would then keep the other study more general--ML, stochastic calc, time series--have the theory and some implementation under the belt would be good.
  • Off-summer internships: now, I don’t feel I am ready for QR interviews and there are not many positions available. The chance of getting a summer internship is bleak. Or is it just my illusion and there are actually some available and I should apply? Are there many internship opportunities in the off-summer season?
I have seen people with winter internship experience from reputable competitors in our hiring pool, but we don't do it and have no idea how that works. If you can find one, I think it adds some value though the duration is shorter than summer internships.

  • Other internships: If I can’t find a QR internship soon, what’s the next best internship option I should try? SWE or data science? Which one would better prepare me for future quant job applications?
We look at machine learning engineer profiles for our QR pool (both internship and experienced hire). Normal SWE and data science backgrounds are usually not considered. I think there is just so much heterogeneity in the DS degree and job role that it makes it hard to consistently evaluate.

  • Projects: if I can’t find an internship in any of these areas, I would like to work on a project with a professor, which topics do you think would make me stand out?
I would tailor it to the expertise of the professor. I think some high quality research should help your case and a good professor should at least make sure it is academically sound. The problem with some of the academic research is that it is often detached from reality and how the market works, and I would especially recommend avoiding doing some sort of theoretical asset pricing research (e.g. stochastic discount factor related), though empirical asset pricing (e.g. predict returns with ML) is ok but a bit overdone.
  • internship after graduation: is it okay to look for internships after graduation? Like I graduate in May in 2024 and do an internship during summer 2024? Or should I apply for 2024 full time positions?
I've also seen this but much more so for undergrads who have graduate school plans. If I were you, I would apply for both the internship and full time.
  • Comparative advantages and niche: since you mentioned before, candidates should have strong core competency, comparative advantages and niche. That’s really enlightening so THANK YOU so much. I am wondering what my comparative advantages and niche are / could be. My comparative advantage could be my statistical / math knowledge but I don’t have much research experience like some PhD have their theses on ML or statistics. My niche might be experimental physics? But I don’t know know how that is exactly related to quantitative finance and could differentiate me from other PhDs in this field. Could you gave me some advice?
Based on your background, it's not something that we typically see in our PhD pools. I'll take this conversation into DM's and ask you a few more questions and address it there.
 
1) Will completing a MFE make up for the low undergrad gpa?
I think having done through the PhD and doing some meaningful research is more than enough to make up for the low GPA if a rational person were to review your profile. However, since a bot or HR might be involved, I would recommend to the best of your abilities, do not make your undergrad GPA available unless absolutely requested. And if they are going to penalize you for it, having an additional MFE will not move the needle much. If you had to provide your GPA and you have some extenuating circumstances, I would suggest writing a cover letter dedicating a section to explain it. Might not get read, but the impact should be strictly non-negative.

2) Based on your experience, what are some realistic starting salaries for quant researchers? Or what google says is accurate (~150k)?
This can vary pretty wildly depending on the firm. In general if you are looking for larger total comp, then try to look for things that are closer to the revenue generation--not all QR roles at all firms are created equal. Being a QR on hedge funds/prop shops/HFTs are most likely going to give you better numbers than in a bank, and someone working in alpha research is going to make more than risk modeling.

I don't want to give specific numbers for us, and there are a lot of forums out there that talk about how some of the biggest packages look for firms within our bracket. In order to compete, we can't really be a small fraction of those and so what we are seeing in general is that new analyst comp have been going up across the industry over the last several years. Of course, it's also much more competitive to get into the highest paying ones.

3) Any advice that you could give?
Based on your questions, it seems that you have concerns over near term finances. It's not a small investment, and since you are doing a PhD in maths, I don't think the degree itself will be too additive. If you want to get exposure to the course work then try to take something at your university. I don't imagine most of the material of the MFE is not that much of a leap to you, and the most beneficial thing might be forcing you to sit down and write the code. If you have other calculus in terms of the invest/risk/returns of the MFE, feel free to reply and I'll try to take that into account in a more elaborate suggestion.
 
(sorry for the delay, was traveling)

I would put some emphasis on implementation (data structure, algorithms, etc). Usually PhD are given a bit of benefit of doubt, but not being able to implement will be a blemish on the interviews, and will hobble you for the actual internship. At some point before you apply, you might want to learn a bit more about the target companies and the kind of questions that they ask. The breadth of knowledge, if you wanted to cover everything, is very vast. I would then keep the other study more general--ML, stochastic calc, time series--have the theory and some implementation under the belt would be good.

I have seen people with winter internship experience from reputable competitors in our hiring pool, but we don't do it and have no idea how that works. If you can find one, I think it adds some value though the duration is shorter than summer internships.


We look at machine learning engineer profiles for our QR pool (both internship and experienced hire). Normal SWE and data science backgrounds are usually not considered. I think there is just so much heterogeneity in the DS degree and job role that it makes it hard to consistently evaluate.


I would tailor it to the expertise of the professor. I think some high quality research should help your case and a good professor should at least make sure it is academically sound. The problem with some of the academic research is that it is often detached from reality and how the market works, and I would especially recommend avoiding doing some sort of theoretical asset pricing research (e.g. stochastic discount factor related), though empirical asset pricing (e.g. predict returns with ML) is ok but a bit overdone.

I've also seen this but much more so for undergrads who have graduate school plans. If I were you, I would apply for both the internship and full time.

Based on your background, it's not something that we typically see in our PhD pools. I'll take this conversation into DM's and ask you a few more questions and address it there.
I must thank you for being a tremendous help when I needed it. Thank you so much for sharing of information and giving me your advice!
 
I think having done through the PhD and doing some meaningful research is more than enough to make up for the low GPA if a rational person were to review your profile. However, since a bot or HR might be involved, I would recommend to the best of your abilities, do not make your undergrad GPA available unless absolutely requested. And if they are going to penalize you for it, having an additional MFE will not move the needle much. If you had to provide your GPA and you have some extenuating circumstances, I would suggest writing a cover letter dedicating a section to explain it. Might not get read, but the impact should be strictly non-negative.


This can vary pretty wildly depending on the firm. In general if you are looking for larger total comp, then try to look for things that are closer to the revenue generation--not all QR roles at all firms are created equal. Being a QR on hedge funds/prop shops/HFTs are most likely going to give you better numbers than in a bank, and someone working in alpha research is going to make more than risk modeling.

I don't want to give specific numbers for us, and there are a lot of forums out there that talk about how some of the biggest packages look for firms within our bracket. In order to compete, we can't really be a small fraction of those and so what we are seeing in general is that new analyst comp have been going up across the industry over the last several years. Of course, it's also much more competitive to get into the highest paying ones.


Based on your questions, it seems that you have concerns over near term finances. It's not a small investment, and since you are doing a PhD in maths, I don't think the degree itself will be too additive. If you want to get exposure to the course work then try to take something at your university. I don't imagine most of the material of the MFE is not that much of a leap to you, and the most beneficial thing might be forcing you to sit down and write the code. If you have other calculus in terms of the invest/risk/returns of the MFE, feel free to reply and I'll try to take that into account in a more elaborate suggestion.
Thank you, I'm planning to follow the C++ course on QuantNet to improve my coding skills this summer.

In terms of risk/return of a MFE, my biggest concern is not able to land a job in finance after a MFE due to my undergraduate GPA. If that happens, I'll probably going to pursue a postdoc while still having loans to pay off. Which is not an ideal situation for me.
 
Last edited:
Back
Top