In the inaugural version of our Interview series - Amazing People in Data

Meet Vladyslav: who started his career building software products, is an engineer at heart, and is now building data solutions in his role as CTO of Mighty Digital. We had a candid chat with him on his journey from the software world into the data world, which is truly fascinating.

Fun fact about Vladysav - He occasionally does some open-source development as a hobby and built a very cool tool called nexrender, for data-driven Adobe automation.

In this super candid and fun conversation, we took a deep dive into his career trajectory and how he entered the data world. If you’re building your data infrastructure, Vlad has some interesting advice for you!

Vlad, tell us about your background. How and when did your journey to the data world start?

I completed my higher education in software engineering and went on to work in the web technology field. And simultaneously, I was doing some game development as a hobby. At some point in time, while I was working actively on my first or second job, I got connected with one of our founders (at Mighty Digital), and we kicked off our first startup. It was a cloth exchange service focused on women's fashion where I was doing primary web-related things and was responsible for most of the engineering and data collection. Data-world then was not as prevalent as it is right now. Mostly everything relied on some sort of general analytics tool. Eventually, our startup unfortunately failed, but we learned a lot from that experience. We went different ways, and then after about three or four years, we collaborated again on another startup. It was called Attractor and was more around data and analytics. So, my primary work and the start-up work were happening simultaneously. And that is how I started moving from software engineering to software engineering + data, where I built data solutions as a software engineer.

Tell us a little bit about your work at Mighty Digital and the kind of projects you’re working on

Most of the time, we work on product-oriented and consumer-facing projects, usually B2C apps. For example - one of our clients is a transportation app that allows people to travel on toll roads in the US. So there are paid roads, and you could either go all the way by attaching a sticker on your front window, or you could use the app and simplify your life. Another example is a scooter company that works internationally and manufactures their scooters and gives people an app for riding scooters. To sum up: we tend to work with startups and companies that are quickly growing and require help with that process, especially in data analytics and data infrastructure. Their growth shouldn’t be bottlenecked by the data ecosystem that they have at the moment.

There is a recent school of thought that is going around in the community around productisation of data. A lot of software engineering principles and development principles are being applied to that data domain. What are your thoughts on this? 

That's a great question. I think DBT is one of the primary examples of the new age shifts that started happening in the data field. Because before that, data was its closed field where you would have these data and analytics teams working in closed-off bubbles in a request-response fashion and only providing limited reports to other groups. With this approach, there is an interconnection between different worlds, i.e., data with software engineering. And as a software engineer, I think that's the greatest thing that could happen to data. This new mix of things improves everything because software engineering principles are strong. They have a whole history and were created for a reason. And if they are applied here, the situation will improve a lot. 

You recently posted your data stack on MDS.xyz. What was starting point of it and how it has evolved over time?

I think this process is entirely organic. First, you look at the tools to analyze the data, like Amplitude, for example. Then you think about how to integrate this data collection into apps. For that, we can use something like Segment or mParticle. So these were the initial tools. After this basic setup, we start looking at much more advanced data integration, connecting multiple sources, etc. That is the point at which a data warehouse comes to play. So you would choose something like BigQuery, Redshift, Snowflake, or maybe even Firebolt. And then, when it comes to data visualization, you can go with a pre-built data studio,  Redash, Looker, Tableau, or something similar. At this point, you will start asking questions like, how do I enrich the data. You either go by building manual pipelines, where Meltano or AWS glue or something like that comes in, or you use Hevo data or Fivetran for data acquisition. Then you ask how I model everything together to get a nice picture? For that, you get dbt. So it’s a process that you discover over a while.

How long did it take for you to build this whole architecture?

It differs from project to project. For me, it was about a year or so. Not all projects require modeling, or not all projects require data acquisition. So, various things came at different times. Moving one step at a time is a more rewarding experience because you're building everything brick by brick. 

What key metrics do you track indicating that the data infrastructure you have built is delivering value?

The primary metric for any data stack should be “the trust” in data. Because usually, whenever a project comes to us, people often say, “oh, is the data accurate?” And that's the primary problem that prevents data-oriented growth within companies. Because if people don't believe that data, they cannot make decisions based on it. The first thing we need to solve is trust, and we need to ensure that data quality is intact. 

We also are seeing an explosion of tools in the modern data stack. There are categories that are getting created every day. How do you feel about that?

People are creative. New tools wouldn’t get created if there were no need for them. There will always be new problems that we didn't see before, and new tools would also be required to solve them. Not every tool and technology will live for long, but people will use it if it solves a problem. And I think it’s excellent!

What would be your advice for companies who are very early in terms of their data maturity? 

It is a good question and the hard one at the same time. I think the best advice would be: to get someone who knows what they're doing. It is imperative. It does not matter if you want to hire an agency that will do everything for you or hire an in-house expert to build everything internally. You would need someone who has experience building systems like that. Otherwise, more often than not, you will lose a lot of time and resources. So, I would say getting the right expertise early on is very important.

You have written somewhere that “more and more businesses will rely on self-hosted solutions for data working and analysis” Don't you think that would be a very big task for a small company in terms of the whole cost and building those engineering capabilities to maintain those systems?

I think it's more about general trends. More and more data-oriented legislation, like GDPR or Ad regulations from companies like Apple or Google, are coming to play. So, we'll see a more considerable shift towards anonymization of data. Companies are worried about how and who is storing their data. And they'll be pushing more towards storing the data themselves. This way, they would have complete control of their data. For small businesses, I would start small, use something cheap and then scale up and migrate to something more advanced.

What are the challenges about the current data space? If you were to start a data company, what problem would you be solving.

We always look for new opportunities to build something incredible. Unfortunately, I haven't figured out something like that. Even though we have gathered many unique problems already, there are often more solutions to each situation. And ultimately, It's great because it saves you time. 

Rapid Fire Questions

What is your favourite tool in the whole entire data engineering space? 

dbt

What is a go-to place to learn about all things, data? Like what, any blog or newsletter publication that you read recently?

Towards Data Science and MDS Newsletters/Blog.

One thing that you like and hate about your job?

Novelty. I love that there is always something new to learn and scope for growth. And the worst part of the data world would probably be data quality! ;)

Very lastly you are based out of Ukraine. How have you been managing all of this and this entire case, how have you been holding it?

We are truly surprised by the world's support, and how our defenders fight for our country is impressive. So, I'm genuinely amazed by everyone worldwide and the Ukrainian Armed Forces. What they're doing is a miracle, allowing us to continue working and helping our economy. And we're happy to do that. We're still in Ukraine. We're still here. So even though some like rockets fly in from time to time, it's okay. We are here, and we're going to continue to be here.

You guys are heroes. I think history will remember you guys for a long, long, long time. There would be stories and fables written.

Thank You!

Check out Mighty digital stack here