Do-It-Yourself Quant student

[finssite]

Hello All,

Please provide me guidance on how to structure my pursuit as a Do-It-Yourself Quant student for the next one year towards a high chance of finding a job? I've put together a list that I need to study for working in a Quant role:

FINANCE
1. Financial markets
2. Financial Instruments
3. Micro & Macro Economics
4. Modeling and Analysis
5. Risk Management

MATH
1. Linear Algebra
2. Calculus
3. Probability
4. Numerical Methods
5. Statistics

COMPUTER THEORY & PROGRAMMING
1. Data Structures and Algorithms
2. C++
3. VBA for Excel
4. JAVA
5. SQL

SYSTEM ARCHITECTURE
1. Fast Speeds
2. High Volumes
3. Multi-Threaded Applications
4. Distributed Systems
5. Using and Integrating 3rd-Party Tools like MATLAB, SAS and others

I understand that solving problems, writing code and working on case studies are essential to building experience towards successful interviews. I'll be taking a practice-based approach for the next one year to develop myself towards interviews. In this regard, I'm looking for detailed curriculum from both academia and work on above items to help me define my scope and pointers to good exercises available online or in books along with their solutions to help me gain close to real-world insights. I'm very glad that QuantNetwork members post very useful information including their on-the-job experiences. For example, a PDF on Quantitative Research QuantNetwork - Financial Engineering Forum and Andy's blog on his work experience at GS have helped me a lot besides all the other posts.

I already have an undergrad Computer Science from Georgia Tech and that I've been doing software design and development for the past eight years including college. I don't intend to become a know-all in any of the above but in some cases I should become excellent and that very good, good and fairly knowledgeable in others. I did perfectly in undergrad Math, Financial Accounting and Macroeconomics and have widely read on Finance. I've applied VBA for Excel and SQL for Financial Modeling, Analysis and Planning applications.

Please let me know whatever comes to your mind after reading the above. I've read a lot that a Masters or Ph.D. degree is considered minimum requirement to work as a Quant so how can I exactly improve my odds without one?

Thank You!
Shakti

 

[Yuriy]

I'm not sure if it is possible to be "do-it-yourself-quant" without strong interest in finance. If you don't know how to do it yourself, you need to go to school
"Do-it-yourself-quant" is possible if you really feel that you want to do something in particular, for example, some people like watching stocks, reading WSJ and they make it their hobby and after that they know what they want to study, or if you have been around finance long enough to answer your question

 

[finssite]

I'm not sure if it is possible to be "do-it-yourself-quant" without strong interest in finance. If you don't know how to do it yourself, you need to go to school
"Do-it-yourself-quant" is possible if you really feel that you want to do something in particular, for example, some people like watching stocks, reading WSJ and they make it their hobby and after that they know what they want to study, or if you have been around finance long enough to answer your question

Hi Yuriy,

Thanks a lot! I've read the WSJ, the Economist and several other sites, magazines, papers and articles over the years. I'm trying to define my scope before starting out for planning effectively. The current list of items define an infinitely broad scope and that I need to narrow down the definitions with specific curricula. After defining the scope, I need to put together a detailed approach for gaining the required skills and that includes the books I must read and the problems I must solve.

I think that college education can fast track the learning process besides many other benefits and that it can take almost two years of college education to attain a comprehensive working set in this area. I can afford only one year of full-time self study so I must focus on the most relevant subset towards a job.

Please let me know your thoughts on the most relevant skills, topics and books to focus on for the next one year.

Thank You,
Shakti

 

[Yuriy]

Hi Yuriy,

Thanks a lot! I've read the WSJ, the Economist and several other sites, magazines, papers and articles over the years. I'm trying to define my scope before starting out for planning effectively. The current list of items define an infinitely broad scope and that I need to narrow down the definitions with specific curricula. After defining the scope, I need to put together a detailed approach for gaining the required skills and that includes the books I must read and the problems I must solve.

I think that college education can fast track the learning process besides many other benefits and that it can take almost two years of college education to attain a comprehensive working set in this area. I can afford only one year of full-time self study so I must focus on the most relevant subset towards a job.

Please let me know your thoughts on the most relevant skills, topics and books to focus on for the next one year.

Thank You,
Shakti

There are two to three books that everybody in quant finance must be familiar with. The famous John Hull's book and the two-volume set by Steven Shreve. John Hull's book was my first introduction to finance back in the Spring of 2002. You need to know most of the topics from the Hull's book. Read that and see if you prefer some topics over others. I personally have my favorites and prefer not to talk about some others

Regarding skills. You definitely need to know C++, statistical or mathematical packages could be helpful depending on where you want to go. You need to know probability, analysis, differential equations, statistics, linear algebra -- most of what students learn in ther undergraduate math curriculum and beginning of graduate.

 

[dstefan]

read the Paul&Dominic guide to getting a quant job - it has a rather detailed study program.

 

[DominiConnor]

It's more detailed now

You seem to be headed in a quant developer direction.

I'd just refine the "speed" bit.
Latency is an issue, which interacts with speed, but is a different thing.

Various forms of numerical analysis like quadrature, monte carlo, finite difference et al will also help your case.

A certain % of employers actually prefer self taught, though there is zero chance of you finding them, without help because they officially say they require paper qualifications. One even specifically has told us they don't want Masters graduates from anywhere, but will take self taught because the head of the group got in that way.

Be aware that a lot of filters work on qualifications.

 

[finssite]

Latency and other thoughts and questions

It's more detailed now

You seem to be headed in a quant developer direction.

I'd just refine the "speed" bit.
Latency is an issue, which interacts with speed, but is a different thing.

Various forms of numerical analysis like quadrature, monte carlo, finite difference et al will also help your case.

A certain % of employers actually prefer self taught, though there is zero chance of you finding them, without help because they officially say they require paper qualifications. One even specifically has told us they don't want Masters graduates from anywhere, but will take self taught because the head of the group got in that way.

Be aware that a lot of filters work on qualifications.

Hi Dominic,

Thank You for a very helpful and insightful response.

I'm going to share some of my thoughts and please let me know whatever comes to your mind on them.

I've read a bit on communication, computational and computer hardware latency and also found out about latency in simulation. These seem to be our high-level options against it:
1. We try reducing it to the longest event.
2. We try hiding it with innovative coding.
3. We're limited by natural or technological barriers and continue with fundamental research and development.

I found out that we need to consider latency for discrete-time algorithms. Could you please point me to some relevant literature at the practical level?

Most business applications are concerned about efficiency which they try to achieve by system design, code optimization and hardware improvement and that latency is generally not considered a critical issue. However, a lot becomes extremely twisted, complex, big and remote when building quantitative applications and that latency becomes a critical consideration primarily for making it all work together in time or even real-time?

I glanced through a paper titled "On the Design and Analysis of Irregular Algorithms on the Cell Processor: A Case Study of List Ranking" and should be reading it just to start learning the way of thinking about complex algorithms. I mention it because of its potential relevance to (2) above in terms of a generic work partitioning technique to hide memory latency.

I'm reading this paper titled "From discrete-time models to continuous-time, asynchronous models of financial markets":
751099035

What is more important in practice, the price (discrete) or the agent-activity (continuous), to analyze the markets? Does this question even makes sense? I ask because this can help better define my focus.

I'll soon read your book on careers in this area. For now, I plan to become a quant developer first to learn the machinery and then move to modeling by acquiring the necessary skills. Later I also want to contribute to sales and trading efforts for eventually gaining a comprehensive experience towards doing more fundamental work in this area.

Thank You very much.
Shakti