Blog | A year ago | 6 — 8 mins

Could a beer start your journey to data science?

Meet Max, a data science lead at Kaskada, and get insights on statistics for beer as well as some advice for life.

Let’s start out with some fun questions. What is your favorite food or beverage?

My favorite food is always sausage pancakes. My family would get together on Saturdays and have a big extended breakfast with my grandfather making them. Then, eventually, I took over as the host. So it's really a family tradition, a family kind of connection, that's why I love them.

My favorite beverage is Guinness, that’s partly due to William Sealy Gosset who is big in the statistics community. Where he's better known by his pseudonym Student. While working for Guinness, he developed a lot of foundational statistical practices for dealing with small sample sizes.

It's really cool that he was one of the first applied statisticians where he was brewing Guinness, which is still around today.

Do you have a preference for on tap vs nitro can or glass bottle?

Oh man. I mean, for me, it's you know, all Guinness is good. It's more of the intellectual idea that actually statistics is useful and that you could like doing statistics. And at the time they were forward thinking, being like let's hire somebody to make our stuff better.

That's cool. Do most breweries do that now, or no?

Hah, I assume, right. He was mostly involved in the agricultural side at the time so he was figuring out how to improve crop yields and stuff like that. Although he did later on become the brew master and head brewmaster of Guinness. Which is also kind of a cool: William Sealy Gosset, statistician - brewmaster. And it was actually a trade secret of Guinness, that they had a statistician on staff and that's why he filed under the name of Student.

What do you like to do for fun?

I'm really interested in game design. Frankly speaking, I approach game design a lot like model design, in that it's all just mathematical systems that elicit certain behavior and responses. On the side, I've built and developed games and I'm interested in how those systems work. Really, it's just all math.

Do you still, now that you're making games, do you enjoy the playing of them or do you enjoy the making?

I like playing games to see how they're constructed. There are two or three main things. There are games that I enjoy seeing how they are constructed. How do the underlying mechanics work? How does this mathematically work underneath the hood? And I enjoy games as storytelling. So if there's a good, compelling, story or well-written story, I'm interested in that. That's just good storytelling. Last there's also the social aspect. So I will often turn off my math side of the brain, in the social contexts, as much as I can.

For example, I will sometimes choose non-optimal choices. Like this is not the mathematically optimal choice, but this seems interesting and quirky. Let’s go for it, this could be funny. I look for something that seems interesting to explore and I like the challenge of kind of breaking the systems.

Very cool, I relate to that. What’s something that most people wouldn’t know about you?

I’m a third degree black belt in Iaido, which is a Japanese sword martial art, awarded from the Japan Kendo Federation. And I've spent a good chunk of time doing that, and tested in Japan three times. I'm relapsed now, but it's still part of my self identity, I see myself as a swordsman.

Did you use a live edge sword like cutting tatami?

Iaido itself generally uses a dull blade, because it’s safer. You don’t generally use live steel until you get higher up. I've used live steel before and it's a very different experience, but mostly I practice with a dull metal blade.

I've cut tatami before, and it's hard not in the way that you think it is. It’s much more about good technique and relaxation than actual muscle, you’ve got to get out of the way and let the sword do the work.

What advice would you give to aspiring data scientists?

Focus on thought processes and analytical processes more so than tools. The tools are constantly changing and evolving. But having a good, solid thought process and foundation of how to approach problems and how to think about problems will absolutely help you and carry you forward.

When I’m hiring, I care much less about what tools somebody knows, because those can be learned more quickly. The world of tools changes every five years. It's completely different. So focus more on those kinds of foundations and how to approach problems.

What was your path to joining Kaskada?

At previous companies I've hit the struggle between the data system that I wanted to have as a data scientist and the data systems I actually have as a data scientist. Kaskada is working to fill a need that I've had at the last several jobs. The solution was either trying to build it myself, and I'm not really a data engineer, so I'd much rather help the data engineers here and work with other smart people to build it for all data scientists.

How did you get into data science?

Well, when I was wee lad of five...

No, joking aside I've always been using math and trying to solve problems and break things down. Going to college, it was kind of a gimme that I would be involved in mathematical modeling. I didn't realize that it was data science at the time, but it was always statistics, applied math, any kind of mathematical modeling and mathematical systems.

And then getting out into the world and realizing that there was a larger community developing around this thing that I'd been working on doing anyways. I had been building mathematical models and mathematical statistical simulations all throughout college.

So then it was like, wait a second. I can get paid to do this. Awesome. I'm going to do this anyway. You know, oftentimes I would play games with math just for fun. One of my favorite “games” was actually just a logistics simulation that I did for a class, entirely in Excel.

Here’s a forecast of what's coming in, here's what the cost function of how we build things. How do we win? And so playing through that, data science is just the newest way to call it. Right? It's statistics, you know, for the modern world.

What excites you the most about Kaskada?

Honestly, I'm going to cheat and give you two things. The fact that we're building a platform that I myself am actually excited to use. It's always great to be able to actually build something that you’re wanting to use yourself.

The other part is the domain specific language for data scientists. I think that the tools that are out there for data scientists don't reflect how we approach problems and how we think about them. Having the opportunity to build a domain specific language and platform that actually helps that are tailor made towards data scientists is great. It’s something that I’ve wanted to do at prior companies, but have never been able to.

My first language that I loved was R because it was a language built by statisticians for statisticians. Sure, it doesn't work as well in certain areas. But I still love it. And there are things about Python and Python implementations to this day where it's like, no, I don’t want to parameterize my distributions like this - that’s not how a statistician thinks about things at all. So being able to actually influence things that way is exciting.

A lot of the data science stuff that exists in Python feels like it was built for engineers and not for data scientists. Right? Where it's like, okay, let's build each of these pipeline things. And then we'll kind of put them in this structure and then connect them with this structure. And you're just like, what the hell am I reading? Where you spend more time focused on the construction of the pipeline and less time focused on the data.

What’s your favorite core value and how do you live it?

Continuous growth and continuous improvement. This actually comes a lot from my martial arts background, a zen koan question that my sensei would ask that I still work to figure out a better answer to every day: How do you beat the man in the mirror?

As long as you're growing and as long as you're improving, you're continuing to thrive. That continuous improvement, continuous development is what I try to live by. Be better than you were yesterday.

That’s a wrap!

Check out some of the blogs that Max has written for Kaskada here (insert link to insights hub blog section), and follow him on LinkedIn where he shares great articles on data science!


Written by Charna Parkey, Ph.D., Data science lead