ULTIMATE GUIDES

The Ultimate Guide to Thriving in Data Science: Quantitative Researcher

November 15, 2017

For many with backgrounds in statistics, mathematics, computer science, engineering, or economics, it’s not about if you’ll make an impact in data science, but where and how. We believe the financial markets offer us unique opportunities to cultivate our skills in a manner that allows us to reach our full career potential.  Whether you’re developing trading strategies that have a tangible impact on the economy or programming the software that brings investment decisions to fruition, you have the opportunity to interact with complex data sets while working with some of the brightest minds in data science. Along with this, perhaps one of the biggest advantages of exploring data through a career at a firm like Citadel is that you receive instant, objective feedback from the markets and can continuously evaluate your work and professional development. Navneet Arora, Managing Director of Global Quantitative Strategies at Citadel, describes data science at a hedge fund as “surfacing patterns that are buried deeply in datasets for the purposes of either identifying causation or making a prediction.”

We want you to make the best career decision based on your aspirations and preferences. We created the Ultimate Guide to Thriving in Data Science to illustrate the types of roles that you could make the most of your skills in – related to the financial markets – and outline how you can reach extraordinary heights over time.

At Citadel, data science skills are applied in three distinct areas: quantitative research, software engineering, and trading. Whichever path you choose, you’ll have the opportunity to simultaneously make an impact on your career and the financial markets. From the importance of collaborating with other teams to always staying inquisitive, find out what skills – across all of these roles – are necessary to reach the market’s summit as a data scientist.

Chapter One of Citadel’s Ultimate Guide to Thriving in Data Science is dedicated to illustrating the role and winning skills of a quantitative researcher. By applying advanced statistical analysis to market opportunities, quants develop and implement next generation valuation models and highly automated trading strategies. In this chapter, you’ll learn what’s required to be a successful quant, with advice from quants who have made an impact and the HR leaders who dedicate their time to helping them succeed. Read on to find out if you fit the mold of a quantitative researcher. If you’re more interested in software engineering or trading, feel free to skip ahead to those chapters.

A Day in the Life of a Quantitative Researcher

Similar to the way a cartographer digests seemingly incomprehensible data into useable information and makes a map of the terrain, quantitative researchers at Citadel are expected to glean insights from datasets and help chart paths to make investment decisions.  While specific duties may vary depending on what team you are on or what type of security you focus on, organizing and analyzing large amounts of data to identify causal relationships in the market is crucial to thriving as a quantitative researcher.

Navneet Arora goes on to summarize what data scientists do in the role of a quantitative researcher: “A quantitative researcher’s role is to blend structured and unstructured data with deep market insights. The objective is to create proprietary trading algorithms. These algorithms help us map out security selection, market timing, portfolio construction and risk management.”

“A quantitative researcher’s role is to blend structured and unstructured data with deep market insights.”

When quantitative researchers are faced with a complex problem, there is a general progression they will follow. The first step any quant takes is always to formulate a theory or hypothesis. Through the development of hypotheses, quants get to continually ask “is this how the world works?” Next, they must identify specific questions that are important to either proving or disproving their hypothesis.  Then, they seek out and identify relevant datasets that might help answer those questions.  After identifying the correct datasets to utilize, quantitative researches will blend structured and unstructured data with market insights and perform statistical analyses to form conclusions.  Finally, they apply their conclusions to create proprietary trading algorithms.

Other responsibilities held by quants include:

  • Processing data and drawing conclusions grounded in logical, mathematical rationale
  • Back testing and implementing trading models and signals in a live trading environment
  • Developing next generation models and analytics

As a quantitative researcher, every day is different, so adaptability and a willingness to confront and embrace unfamiliar or uncomfortable scenarios is another key to success.

Hard Skills – The Essentials

The work of a quantitative researcher is mission critical to firms like ours that work with some of the largest and most complex datasets – financial markets.  Accordingly, hedge funds have high expectations for individuals in these roles.  Here are some of the skills a strong “map maker” brings to the table:

  • Advanced knowledge of probability and statistics
  • The ability to reduce data to its mathematical core
  • The aptitude to identify patterns and trends in data sets
  • The deftness to formulate hypotheses and actionable solutions
  • A baseline understanding of how to translate algorithms into code
    • Some of the most common programming languages for quantitative researchers at Citadel include C++, Python, MATLAB, and R.

In addition, it is just as important to understand the companies and sectors you’re analyzing. As Brad Lau, Deputy Chief Operating Officer for Surveyor Capital at Citadel, observes, “data scientists have to understand the companies behind the data. What are the actual drivers of the company?  The conclusions are only helpful if they can be incorporated into an investment process.”

Soft Skills – Tricks of the Trade / Pro Tips

The reason that so few can call themselves the best of the best in quantitative research is that excellence in this profession requires a unique combination of technical and “soft skills.” Some of these soft skills that help the best stand out include:

Collaboration

One of the most important soft skills that a quantitative researcher can cultivate is the ability to  develop deep relationships with peers and other teams.  A quantitative researcher often has to communicate conclusions to both traders and software engineers in their terms. Brad Lau puts it this way – “A cartographer designs a map so that the rest of us can follow it effectively. Quantitative researchers must communicate their conclusions in a manner that allows the rest of us to make effective investment decisions.” These conclusions will need to get translated into tangible inputs into valuation models and trading strategies.

Curiosity

To be a successful quant, you must have an understanding and acceptance that you need to probe at every step along the journey to solve a complex problem, from defining the problem to accepting a conclusion.  We seek to understand the What and the Why.  According to Navneet Arora, the well-trained data scientist knows to look at any of their findings or conclusions with some degree of skepticism. Finding the answer is only the beginning. You then question the answer.

Humility

You will undoubtedly want to prove your idea or hypothesis. However, you will find that sometimes, the result of your analysis disproves your hypothesis.  You must remember the goal is to learn and seek to understand truth.  Do not become discouraged when the result of your research is not what you predicted.

Courage

According to Candice Berger, Co-Head of Campus Recruiting at Citadel, to be a successful quantitative researcher, “you have to be someone who has the courage to navigate ambiguity.”

Advice

One of Navneet’s tips to current students and aspiring data scientists is to take the initiative to apply the foundational knowledge gained through your classwork to real-world scenarios.  There’s a wealth of data available, so pick an empirical problem and try to solve it using the following steps:

  • Design how you’ll collect data
  • Choose the correct methods for analysis
  • Derive conclusions
  • Present the solution to a friend
  • Critically analyze the results

When trying to solve these problems, Navneet recommends that you “treat data like clues and not conclusions in and of itself. You want to see the whole map, not just the one point in front of you.”

As with all data science problems, before you start compiling data and performing analysis, you should have an expectation for the result, and keep your final objective in mind. When attacking a problem for the first time, remember to utilize your resources effectively.  A successful data scientist knows which tools to use for which jobs.  According to Navneet, “the best data scientists use different tools for different problems. It’s almost like on a golf course, you have to figure out which golf club are you going to use for which shot.” Some of the tools that data scientists learn to wield and apply include:

  • Regression analysis
  • Fundamental analysis
  • Machine-learning

Navneet also tell us that at Citadel, “up and coming data scientists are paired with experienced data scientists to learn about the contexts under which these tools should be applied.”  We recommend that you consistently ask managers and mentors about whether your chosen tools and methods are applicable for the problem you’re trying to solve for.

While your coding abilities are not expected to be equal to those of a software engineer, it’s recommended that if you’re still in school, to take as many computer science classes that are offered.  These skills will be applicable to your work as a quantitative researcher.

Finally, you need to be curious and learn about the business and goals from leaders. You should attend conferences and competitions to both learn from and network with industry leaders. For example, Citadel hosts The Data Open, a datathon competition taking place throughout the year at a number of top universities. These types of competitions and conferences can also be a great way to stay ahead of the curve, as data science is a rapidly changing and fast-moving field.

The Citadel Approach

At Citadel, quant researchers are pushed to learn and improve each and every day. In Navneet’s words, “at a hedge fund, I don’t have to wait months or years for feedback from a beta launch. The markets are going to tell me if my hypothesis is correct and thus, trading decisions are impacted immediately.” We aim to provide quantitative researchers with the tools, mentors, and culture to act on this immediate feedback and develop faster than they could elsewhere. We expect our team members to challenge themselves and those around them to take in this feedback and continuously refine their approach to understanding the markets.

In addition to a focus on feedback loops, we aim help quantitative researchers unleash their passion for their profession. According to Navneet, “at Citadel we’re in a constant race to innovate and identify new patterns. The market moves really fast and it’s our job to move even faster. That’s the definition of exhilaration. We match that exhilaration with a passion for understanding the markets.”

We’re committed to making a difference and that difference often shines through during the conversations we have with each other at an event series called After Hours. Team members, including many of our quantitative researchers, get together to socialize, network and reflect.

We’ll discuss how capital, effectively deployed, can help lead to new health care products moving to the development phase, or allow a pension fund to support retirees.

Chapter Two: Software Engineering

Chapter Three: Trading