The question of how to become a quant is asked very often by current students and working professionals looking to make a career shift. This post will serve to answer a lot of common questions and points about the path to becoming one. Before I begin this first blog post, keep in mind that this is only my personal view. To understand this topic, I recommend reading many different points of view on this and completely expect that some people will have differing points of view, especially when it comes to things like what kind and how much education someone ideally should have for these kinds of roles. That being said, I’m hoping I can clarify a few things that I’ve learned over the years in terms of how one can become a quant.

Defining a Quant

One aspect of this question that makes it tough to answer is the definition of a quant. In the 2000s, quants referred usually to someone at a bank or hedge fund that was building quantitative trading algorithms, risk models or creating complex products like mortgage backed securities. Currently, this definition has been used very loosely to define all different types of work. I like to think of the current landscape as having five different types of roles, that are often overlapping. For example, one can be doing both data science and business analytics work and one might even try to say they are the same in a certain scenario. I only separate them because sometimes there are business analytics roles where you find no coding and everything is done through excel and tableau/similar programs for example. The five ways I like to define them are:

1. Business Analytics: Roles where you are providing analytics and insights for business decisions utilizing programs like tableau and excel. A role exploring costs and revenues within an organization and visualizing it for senior management is a good example.
2. Data Science: Roles where the employee utilizes statistical programming languages such as Python, R and Matlab and possibly languages such as Java, C++, etc. to clean data, draw insights, visualize them and build models. Very often machine learning is a core part of the job and some may be focused on using Hadoop or Spark to work with big data (the kind of data you can’t fit into excel).
3. Quant: These roles are much more heavily based on model building. While the data scientist might be creating insights in the context of a hedge fund, the quant is responsible for building out trading strategies, applying optimization for things like risk control, and more. They can also be utilizing machine learning depending on the scenario. Quants working on problems related to prediction of default will often utilize some kind of machine learning to build stronger models.
4. Developer: A developer will be writing the libraries and programs at companies. In the context of banks and hedge funds, these are usually called quant developers because they are building out the functionalities that the quant team uses for things like fast querying of databases or implementation of quantitative trading algorithms. In the broader context, there are numerous companies that hire developers for a wide variety of tasks such as software development, web development and more.
5. Technology: Those working in technology are in charge of the actual infrastructure such as the databases and the company hardware. They do not necessarily need the ability to code in languages like Python/R/Java/C++ because they are primarily focused on the actual technology, and the technology specific languages which they need to understand.

It is important to understand which skillsets you want to build. In my experience, you are always going to build more than one but rarely you need all five. Overall, these are just my own loose definitions that I like to use when thinking of the landscape. I will focus on the quant/data science roles, with attention given to the finance versions of them but I will try to as well generalize for those of you who are interested in non-finance jobs.

The Formal Education Question

For a lot of quant roles, you may find a requirement of a master’s degree or a PhD. Sometimes it is only preferred, but for certain there are a lot of roles where an advanced degree is crucial. There are also roles referred to as junior data scientist, data analyst or quantitative business analyst that seem to generally be more entry level and only require a bachelor’s, but this definitely depends by company. Overall though, a lot of quants at top companies will have a master’s or PhD because of how much technical materials go into understanding quantitative models. There are also professors at universities who can be employed as advisors or consultants at companies to help with solutions given their deep domain knowledge.

Another general trend is a preference for generally mathematical or quantitative majors over business or finance majors. You may have noticed, but there are many physics PhDs who now are quants. It might not make sense intuitively, but physics is similar to quant finance in the sense that one must solve complex and mathematical modeling problems. Because of this, majors like computer science, mathematics, physics and engineering are often preferred for quant roles. The idea is that the quantitative aspects of the role can be harder to train than the financial aspects.

The Overall Education Question

Whether or not you opt for an advanced degree, there are still areas that are essential to learn to be proficient. The financial areas are very dependent on the job, but the quantitative aspects tend to be generally applicable across disciplines. Of course, there are quantitative models more applicable to some roles, but often you will see that generally working with data wrangling, visualization, validation/stress-testing and research presentation are all going to be important for each role.

Coding skills and data wrangling skills are a key to success. Learning programming languages such as Python, R, Matlab, C++, Java, etc. will be extremely important. Which languages to focus on are outside of this post’s scope, but you can get a feel for this by looking at job postings. If you find that roles you are interested in keep asking for certain languages, those will be the most beneficial to learn. If you are unable to do college courses in computer science, there are many free resources that help to fill the gap on edX, coursera, udemy and other MOOC providers. These courses can help build the basics. Overall, the best way to truly build the skills is to begin working on projects you are interested in once you have a good basic knowledge. This will force you to really hone your skills. For example, if you want to build trading algorithms, you can begin testing your own trading hypotheses. Even if they are not strong trading strategies, you will be able to work through the process of researching a strategy.

Another important skillset is your mathematical background. Especially in the context of machine learning, it becomes vital to understand the mathematics behind models. The truth is that thanks to plug and play libraries, one could avoid learning the math and very quickly learn to build neural networks, classification models and other advanced models, but this will be dangerous. Not understanding the mathematics behind different algorithms can lead to making poor choices when formulating models and working to clean the data beforehand. The most important courses are probably going to probability/statistics as well as linear algebra since so many of the models rely on these fields. Calculus is also important (to understand how neural networks work you certainly need it). More specified coursework can also be necessary such as Bayesian statistics if you are looking to really understand Bayesian modeling.

The last piece is not always quite as vital, but understanding technology can save hours of time. Some of the most pertinent examples are GPU and cloud computing. GPU (or graphics processing unit) refers to using GPU on a computer for much faster machine learning, which can save huge amounts of time because it runs computations at a speed that normally can’t be attained through regular methods. Cloud computing is another example of technology that can be extremely helpful. When you are running huge simulations, it can be massively helpful to be able to split tasks into multiple virtual machines on the cloud to parallelize and speed up tasks. If you want to run the same code but with 20 different starting parameters, and each run takes an hour, you can either wait 20 hours on your local machine or you can run the code in an hour on 20 virtual machines. While you won’t need to be an expert in AWS or Azure, these technologies are worth investigating for specific use cases.

Getting the Job

The final question is how does one get a job as a quant. The answer is, as in many cases, it depends. There are a few paths that can be helpful though.

The most obvious way is to simply apply to jobs. Some places tend to look for years of experience or advanced degrees but there are certainly also programs that look for entry level quants. Recruiting agencies can be very helpful in this since they will get your resume in front of people with better luck than applying through job posting.

Beginning in a general finance role with projects involving quantitative analysis is one possible way. If you are able to work at a bank or other financial institution where you are allowed to work with some modeling, and also learn the quant skills, it can lead to a more quant focused role down the line.

A final option is generally networking and working on projects/research papers. Through working on tangible projects or networking with people in the industry you may be able to find roles through references from people you meet along the way.