How Travelling and Hiking led Kurt to Automation and Systematized Algorithmic Trading

10 min read

Retail investors have been entering the global trade markets at an astonishing rate in the last few years due to the improvement of accessibility and the decrease of frictional broker fees.

At the same time, there has been a proliferation of data that before had a high price tag to use. Price data, fundamental data, and alternative data are now at the fingertips of both hedge funds and retail traders alike. Raw signal speed is no longer the key winning factor but rather what system can interpret the variety of data signals and be right when it most matters.

Kurtis Selander (Kurt), is a Computer Engineer from California. After graduating from UC San Diego, Kurt partnered to create an Aerial Engineering Consultant Firm that specialized in one-off custom aerial platforms. This gave him a strong foundation in systems, agents, data, and automation before venturing into systematized algorithmic trading after finishing EPAT.

He's currently focusing his efforts on the structural geometries of data and the combination of machine learning approaches. Kurt believes a lot can be achieved by combining unsupervised learning techniques to identify hidden structures and then using reinforcement learning processes to train a neural network for trading has significant potential for many applications. He believes in ‘data-first’ decision-making models.

Kurt shares his journey from UC San Diego to his latest experience with algorithm trading.

Hi Kurt, tell us about yourself

My name is Kurtis Selander and I'm from San Diego in California, US. I have a Masters in Financial Algorithmic Trading and a B.S. in Computer Engineering from the University of California, San Diego. I spend much of my free time hiking mountains and on tropical islands surfing.

My Dad is a Chemical Engineer, and my older brother is from Computer Science, so I essentially just chose a combination of both disciplines when I went to college. I was always interested in technology growing up. My dad always had computer parts strewn about his office as he plucked away on a few different computer builds or rebuilds. He enjoyed playing with the parts - assembling and disassembling them.

As I became further educated on the history of the world and where we were on the timeline, I looked at technology and heard my dad’s stories of how computers used to take up an entire room.

  • I remember my dad using floppy disks, moving to writing on cd-disks, to then using a USB.
  • I remember AOL internet having a terrible sound to connect and completely occupied the phone line.
  • I remember when 56k was fast for internet speeds and how cool it was to instant message with friends or socialize on myspace.

Fast forward 20 years and now nearly every individual on the planet has a small box they keep in their pockets. It is capable of accessing the world’s data at their fingertips anytime and anywhere. Phones, mobile computing, and our data networks have made it possible for individuals to make a trade and participate in the market from the top of a mountain or from a tropical island.

What I eventually came to realize is that by creating an automated system, when I am out on hikes and don’t have internet or when I am on a beach with no internet, the system will automatically execute my rules for me. Exactly how I wanted it, faster than any human could ever process and react by itself.

While on the road, I considered it a success if I could hike a mountain and listen to a book on the same day. So that is what I did, day after day, for months. I listened to all types of books and found a few heroes I wanted to learn from in finance such as Warren Buffett, Peter Lynch, Charlie Munger, David Dodd, Benjamin Graham. I also searched in biographies of Elon Musk, Philip Knight, Steve Jobs, and Bill Gates to get clues on where these visionaries see the future going.

And now on the other side of the QuantInsti course, I’m equipped with a skillset and a solid foundation of investment rules passed down by some of the greatest investors of all time.

How did you get into the world of Algorithmic Trading?

The loss of a couple of close people, created a shift in my value system, launching me into the idea of travelling to South America with a 1-way ticket.

Two years later after a thousand different experiences, challenged me, connected with different cultures, met all walks of life, working different odd jobs. This really helped me improve my perspective and understanding of the world and reality.

Because I had more time on my hands, I was looking for an alternative way to make some more travel money. I stumbled upon blockchain, crypto, and the idea of smart contracts/automation.

The idea of smart contracts really hit a chord with my robotics background and desire since young to make a ‘smart automatic home’. But I had another connection to make between automation and finance that wouldn’t hit me until I was in the mountains hiking without the computer or the internet.

I was now buying and selling Bitcoin and learning about the ecosystem when I would then have no access for 1-2 weeks. For those in the crypto world it, was astonishing how fast up and down the digital asset moved. I’ll never forget the feeling of coming back after 2 weeks to a significantly large drop and thinking to myself ‘how can I protect myself here better next time?

After many YouTube hours, I became familiar with MACD, Stochastics, RSI, and order book structure. I realized that really what I was doing was looking for a stack of patterns or a ‘state’ of patterns which gave me a probability confidence of up, down, or neutral.

These patterns were what I was repeatedly looking for and on my next trip up into the mountains, I thought maybe there is a way to systematically and programmatically combine them to buy on a buy signal and to sell on a sell signal. This was the start of financial algorithms and automation in my python algorithms career.

I tried to hit the ground running by popping into google and searching ‘How to trade stocks programmatically’. What I found did not enthrall me.

There were many either extremely complex, or extremely expensive systems. I wanted to be able to backtest an idea to see its competency and then launch it in the same area.

  • There was the gold standard Quantopian which was great to share ideas, but you could not launch from there and then was eventually bought by a billionaire and shut down. Even if you could get Quantopian to work, it was in a virtual environment that had limited access to python packages.
  • There was the gold standard, Interactive brokers but it seemed overly complicated to get going with the full-fledged pythonic API.
  • Some services gave data that you wanted to process while you might trade with a different broker.

The world of python seemed overly complicated because there are hundreds of important packages and nuances to just hop in, without any structure to approaching the large learning curve.

I realized that what I needed was a school where I could develop a stronger mathematical foundation to help interpret the code. I needed direction in knowing where to start my focus so that I was not studying ineffective material. This brought me to search online and I found EPAT.

It was there that I learned how to get paid data and scrape data from the internet (price, fundamental, alternative), process that data, create alpha signals, backtest the strategy, analysis of the strategy, bet sizing, and portfolio/position risk management.

You’ve worn many hats in your career - Founder, Engineer, Sales, Teacher. How did that come about?

In college, I found that class wasn’t holding my attention very well so I started exploring projects that friends were working on in the research labs on the campus.

One friend of mine was developing drones for ‘The National Geographic Engineers for Exploration’ in order to find Genghis Khan’s tomb over in Mongolia. Needless to say, I thought that was the coolest thing since sliced bread.  We felt like the Wright brothers creating new flying inventions.

We became familiar with systems, system controllers, loops, signal inputs from accelerometers/gyros/barometers/cameras, desired motor and GPS outputs, high-pass filters, low-pass filters, signal interferences, system noise etc. My ‘nerdy’ background extends from here.

Our aerial robotics firm accrued an impressive list of clients with the likes of Twitter, Google, General Electric, Caterpillar, Bloomingdales, and DJI. Fresh out of college, we had to grow up quickly.  Drones were and are a fast-moving industry so it was either sink or swim.

We had to hustle during the day to create more deals, business sales, marketing, and outreach while during the night we would stay late in the office working on our product deliverables.

Other than that experience, I have always had a keen desire to experience and explore the world.

After a best friend passed away due to depression, I realized life is just too short. And so, I packed up my bags and just left for a 1-way ticket adventure which led me to travel the road for over 2 years. Reflecting, this is the time that really allowed me to test and fail and to learn what was interesting to me rather than what a school wanted me to learn.

This period gave me foundational confidence that I can handle whatever comes my way and most likely the confidence that I can understand all the complexities that go into making a successful algorithm.

Along the way… Yes, I’ve even worked as an English Teacher for a few months in Acacias, Colombia!

But my main goal now that I'm working toward is developing myself around machine learning ideas and principles. I really think you can do a lot by using unsupervised learning to find hidden structures in the data. And then you can iteratively simulate like a game using reinforcement learning to find the best route leading to the highest reward function.

How did you overcome the odds to learn algorithmic trading?

I had ideas in my mind that I wanted to implement but was held back by a lack of a programmatic skillset. Programming was never natural to me. Math was. But programming always just seemed a bit hard and there was always too much memorization for each language.

I did not have a strong command of Programming. I started learning Java, Javascript, Ruby, Rails and Django. But I felt like my knowledge of the languages was always just a little clunky.

It actually wasn't until I came across the book “Zen and the Art of Motorcycle Maintenance”. This book gave me the perspective to appreciate what beautiful code means to me and what could be beautiful about a programming language, rather than just kind of looking at it as a means to an end.

What was stopping me from algorithmic programming were:

  • My knowledge base of the Python language.
  • Lack of knowledge of how to backtest my strategy.

It wasn't until I had more experience with languages and the personal persistence of becoming fluent in another language, that when I went back, did I finally start to understand the language more deeply.

I started seeing Python as a beautiful language. A look at the language and I’d say, “Oh my gosh! I can do that!” It's going to take a lot of work, but now I believe that I can do it. I think that belief is just as important.

On my journey to learn python and to automate trades, I stumbled upon Quantopian, Interactive Brokers, and EPAT. Some instances made me realize that I did not have the skill set to understand and execute a trade. I could go on to YouTube and do some self-learning.

But I really cherish the idea that everybody needs mentors. While we can walk that path alone, it is much easier if we know a mentor or a guide, essentially, who has been to the top of the mountain and is sharing some of their knowledge.

This really pushed me towards EPAT. Right off the bat, I just realized that there's a lot of good information here that I've never even thought about. For example, there are many nuances to ensure that our logic is not ‘looking into the future’ when backtesting.

Before using python to systematize my trading, I spent a lot of time on Microsoft Excel and Google Sheets trying to build models manually. So, I have a solid foundation of the structure of data in excel. Seeing the cell-by-cell breakdown of a time series in excel by the EPAT professors was extremely helpful in understanding the nuances of the topic.

Now that I know the algorithmic structure taught by EPAT, I can break up what seems like a monumental task into manageable chunks and steps.

Which feature of EPAT do you like the most?

EPAT has been incredibly useful. I'm feeling a little power. It feels like magic.

It's cool, the feeling that if I have an idea, I can actually make it. There are some very, very useful skill sets that I learned through EPAT.

  • I can manipulate and combine data structures in my head.
  • While programming, I can build using essential tools and python packages.
  • Knowing the end structure of the data object we want to create, I’m able to cut, order, prepare, normalize, select, and combine it with other data into a single informative data structure.
  • I can bring in data from any API. I can bring in data from Yahoo Finance, or IE, or Polygon and all of the different data exchanges.
  • I can scrape any information that is in a structured form on the internet.
  • I can bring in news data, stock, prices, fundamental and technical data all under one area to create alpha signals.
  • I know how to put data together to feed it to a machine learning algorithm.

I'm very happy and stoked about the things that I am going to create in the future with this foundational algo base.

What message would you give to the individuals who wish to pursue algo trading?

I’d like to share the following points that have guided me and I hope will guide you too!

  • Be Persistent!
  • Foster a deep understanding of data - In the end, what we're doing is we have panel data and we're just trying to find structure and manipulate it.
  • The effort that you put in grows exponentially - Imagine, if you were to figure out some sort of trading algo that can get you an alpha of even 2% or 3% and you use that for 10 years. Where would that put you now?
  • Plan - Write down how much money you need to survive in a month or even a year. And really put it down on paper. Then the question becomes “how can I make it happen” rather than “I CAN make this happen.
  • Ask productive questions that solve the problem at hand - I think that little mental switch about, “how can I find the tools to handle my risk management,” or “how can I find a tool to develop the knowledge and my Python programming?”. This will really help you achieve your goal.
  • Live Life! - Lastly, go outside. Because being healthy and happy and eating right helps your brain solve problems more efficiently. If you don't have that right, then whatever you’re working on is moving slower.

All the best!

Thank you, Kurt, for connecting with us and sharing your journey. It’s great to see someone with such a diverse background take on algorithmic trading. We’re glad to be a part of your journey.

If you too desire to equip yourself with lifelong skills which will always help you in upgrading your trading strategies. With topics such as Statistics & Econometrics, Financial Computing & Technology, Machine Learning, this algo trading course ensures that you are proficient in every skill required to excel in the field of trading. Enroll in EPAT now!

Disclaimer: In order to assist individuals who are considering pursuing a career in algorithmic and quantitative trading, this success story has been collated based on the personal experiences of a student or alumni from QuantInsti’s EPAT programme. Success stories are for illustrative purposes only and are not meant to be used for investment purposes. The results achieved post completion of the EPAT programme may not be uniform for all individuals.

Download EPAT Brochure