The Artists of Data Science

The ONLY self-development podcast for Data Scientists

The Secret to Success Is In This Episode | Kyle McKiou

On this episode of The Artists of Data Science, we get a chance to hear from Kyle McKiou, a data scientist who took the lessons from his own struggles that he faced attempting to break into data science and packaged them into a course for up and coming data scientists. He is known for his remarkable talent for building skilled, balanced and productive teams. He gives insight into how he broke into the data science field, his approach for problem solving, and they importance of facing your fears.

Kyle shares with us the importance of finding a mentor that can guide you to accomplish your goals and the important soft skills that you may be overlooking. Kyle brings unprecedented wisdom and advice to this episode, and the points he outlines can help everyone step up their professional goals.

Some notable segments from the show

[7:43] What value Kyle believes data science will bring within the next few years

[11:38] How to transition into data science
[16:33] The importance of cultivating a growth mindset
[28:30] Soft skills that data science candidates are missing
[33:01] The single biggest myth about breaking into data science

Kyle’s journey into data science

Kyle’s journey began back in college, when he decided to get a PhD in mathematics. His original plan was to work as an investment banker or at a hedge fund. After a year into his classes, he realized that he didn’t want to be a part of the banking system. He wanted to find a career that was more fulfilling.

He eventually decided to drop out of his PhD program, and left with his masters degree. He then got a job as a software engineer. Although the job had aspects that were interesting, he still didn’t feel fully satisfied with the impact he was making in his role.

One day, Kyle stumbled upon a Harvard Business Review article about data science being the “sexiest job of the 21st century”. After reading that article, he decided that he wanted to transition into data science. The field promised a role where he would make a real impact in a business context.

[3:30] “Sure. So I originally went to school because I was interested in working at a big hedge fund or a big investment bank or something as a quants.

So I decided I was gonna get a Ph.D in mathematics. So I made the switch. I was actually studying exercise science at the time. I made the switch, and right into my first semester of college at University of Illinois, I was taking senior level math classes. I took five math classes my first semester. So I was working my way through the mathematics degree. And then after a year or so, I realized that I didn't want to really be part of the banking system. I didn't really see that necessarily adding a lot of value to society. I didn't think that was going be very fulfilling. So I was kind of stuck in this position where I had a mathematics degree. I'm like, man, I don't know what to do with it because I don't want to work in banking. I don't want to be a math teacher. I don't want to work for the NSA. What can you even do with a math degree? So I started looking at what can I do with this? How can I start applying this knowledge? And that's where I started studying economics and computer science and statistics and all these other fields that were related to math. And I got really interested in doing mathematics on computers. So I ended up starting doing a PhD. in scientific computing. And then I decided that a life in academia wasn't where I wanted to go either, because that's just a lot of research and writing papers.

And you never really get to do anything. You don't really get to make a real impact in the world. You don't really get to apply your knowledge. You just try to find new knowledge for the sake of it. So I ended up dropping out of my PhD program. I was actually doing a computer science PhD at University of Illinois at the time, and I just left with a master's degree. And so I had a masters degree and then I got a job doing software engineering. We were basically helping companies do electromagnetics simulations. So if you wanted to design a stealth fighter or a battleship, we were creating the software for you to do those simulations and development. Now, this sounds really cool, but the problem is that all of our clients were classified. So it's basically all top secret. And we would roll out this new software, all these improvements, and nobody would say a single word back. It's like if you had clients that just totally ignored you and then you thought, well, what the hell? Do people like the product? Is this helpful? Is this useful? Are they enjoying it? Or is this just a big waste of time? So it's a little bit frustrated that I couldn't really tell if I was making an impact. And that's when I learned about data science. I think I probably saw the same article that said that, you know, everyone else off Harvard Business Review, sexiest job of the 21st century. A friend sent it to me, and I looked at it and said, man, this seems cool because you really get to make a real impact on businesses. You're much closer to the business side, the application side and the work that you do, while it's still mathematics and computer science and statistics and these are things, it's actually used by the business to get a real result.

It's not just theory, it's put into practice, and that's how I got into data science. I said "Man, how do I really make a difference, a positive impact in a company that I can see?" And that's why I decided I got to make this transition into data science.”

Where is the field of data science headed in the next 2-5 years?

Data science has been slow to adapt. It needs to become more scalable. The way to accomplish this is by making data science more systematic. It has to become more organized, with an engineering focus, instead of an analytics focus.

Many companies have tried to implement data science practices quickly. They focus on how to make money with the data they have. This does not work, simply because it is not well planned. You cannot scale with this approach. This is why companies struggle year after year.

Creating a systematic engineering focused discipline will make data science scalable, repeatable, and adaptable to new markets. Doing this will allow data science to have a large impact for the company.

[7:43] - “What needs to happen to make data science more scalable is it has to be much more systematic. It has to be much more organized. It has to have much more of an engineering focus and not just an ad hoc analytics focus, because a lot of companies have tried to set up data science practices really quickly.

What they do is they get one person or maybe a couple of people and they say, "oh, we've got some data, what do we do with it?" “How do we make more money with it?” That's just an approach that doesn't work at all. Then it ends up, of course, not working because it's not planned out. There's no real way to get value from this. There's no way to scale it. There's no way to make it repeatable. Then they say, "oh, data science doesn't work for us". They shut it down or they struggle year after year. So really, it has to be a much more systematic engineering focused discipline because that's what makes it scalable. That's what makes it repeatable. That's what makes it adaptable to new markets, to new problems, to new situations. When you can do that, that's when it makes a lot of money and a lot of impact for the company. That's when data science flourishes, gets a bigger budget, makes a bigger impact. So it's really a focus on engineering, more so than just can we build a machine learning model or do we have, “artificial intelligence?”

What will separate great data scientists from the rest of them?

What separates great data scientists from good ones is not dependent on skill in statistics, building statistical models, or even software engineering. It’s being able to understand the context of the problem that they are solving.

You can be a good data scientist if you can build good models, but if you want to be great, you will need to understand the data, where it came from, how it was collected, the relevance of the data to the business, and what problem you are trying to solve. This will allow you to build a model that is specific to your business problem. Too many people want to build models that have high R-squared or have low error, but this doesn't necessarily make the largest business impact.

Your model may not fit the real world, because the model is a hypothetical, perfect situation, while the real world is messy with multiple angles of interpretation. So understanding the context of the problem is what takes a data scientist from good to great.

[9:56] “What will separate the great ones from the good ones is not their skill in statistics or building statistical models or even software engineering, for that matter. It's really understanding the context of the problem that they're solving. So you can be really good if you make this repeatable, if you understand how to build good models. If you can take data and turn it into predictions. But if you want to be great, you have to understand the context of the data, where it came from, how it was collected, the relevance to the business and the problem that you're trying to solve. So you can build a model that's, you know, much more subtle and specific to the situation of the problem you're solving, because a lot of people just see numbers and they say, "oh, well, I want to build a model that has a high R squared or that has a low error", and it's great to have a model with a small error, but that doesn't necessarily make the biggest business impact. So it's realizing that sometimes your data is biased, sometimes your data is not good, and a model that has more accuracy, is it, you know, whatever your accuracy metric is, but a model with more efficacy is not necessarily better in the real world, because, you know, this is a hypothetical, perfect situation model. Whereas the real world is big and it's messy and there's a lot of different angles to it. So understanding the context of the situation is really what takes someone from good degrade and making a moderate impact to a huge impact.”

Key takeaways from the episode

Why you need to do things that scare you

[19:38] Move towards the thing you fear. Fear is an indicator that there is an opportunity for you to grow here.

Why creating a system is the key to successful problem solving

[24:29] Don’t just solve the problem. Solve the problem by creating a system.

[25:49] Too many people stress the importance of problem solving by having an original, creative approach that is very innovative. This can be a daunting task, and it is very difficult to always have this approach. Instead, focus on breaking down your problem into smaller, actionable pieces.

How to work with non-data scientist stakeholders

[31:05]

1.) Understanding what other people want and what other people need to be successful.
2.) Formulate the solution that's going to help them.
3.) Communicate to them, in their words, why and how it is going to help them, because you need their buy-in.

The importance of modeling someone who you want to be like

[42:01] Find someone who has accomplished the goals you want to accomplish, and have reached the levels of success you are aiming for, and take their advice. That will accelerate your growth.

Important Soft Skills:

[29:22] The most important thing is your communications skills and being able to present your ideas.
[29:31] Understand what other people want, then align yourself with their interests.

Memorable quotes

[16:13] “Be risk averse; Test everything.”

[24:50] “You've got to engineer a system that solves the problem for you, because if you have to leverage your own intelligence to solve a problem, well, you're going to be very limited in the amount of work that you can do.”

[27:23] “...you start with the problem you want to solve. You break it down to simpler problems. You break those problems down to simpler problems...all the way back until you get to your present state and then you see the exact path forward at any point time…”

[28:31] “...I think in most careers it's not going to be the hard skills that separate you, particularly in data science…[it’s] those soft skills, because you realize that if you want to make an impact in the company as a scientist, you're going to need other people to work with you…”

[34:55] “...it doesn't matter how much you know, it matters how much you can learn and adapt.”

The one thing that Kyle wants you to learn from his story

[37:12] “ Do the thing, and you will have the power.” - Ralph Waldo Emerson.

The one difference between people that are “successful” versus not successful is they've done the thing. They've done the work. So if you want some sort of result, it really just comes down to you putting the work in to get the results.

Be patient enough to not quit. That's all there is to it. You put in the work and you keep improving and you don't quit. And if you do that, you'll get there. I guarantee it.

Lightning Round

Best advice that Kyle has ever received

“You should find someone who knows more than you and has had some of the success that you're looking for, and take their advice”

Advice that Kyle would give to his younger self:

[39:45] You can do anything you want. You just have to do it.

Books and other media mentioned in this episode

“The 10X Rule” by Grant Cardone.
“Mindset” by Carol Dweck.
“Psycho-Cybernetics” by Maxwell Maltz
“Deep Work: Rules for Focused Success in a Distracted World” by Cal Newport

How you can connect with Kyle

LinkedIn
Instagram
YouTube
Data Science Dream Job - Free Webinar

Article Comments