Yuliana Havryshchuk makes data crunching easier for her colleagues at Zillow. She’s a data software engineer, helping build processes and automated tools that help sort through datasets that her company collects. These tools, in turn, make data easier for analysts and data scientists to parse through, fueling business decisions and giving Zillow’s customers a better experience. Havryshchuk has previously worked at Buildzoom, Hack the North, and Apple, developing her skills as an engineer and developer. Here are her experiences in this big data engineer profile.

Tell me about yourself.

I am currently a data engineer at Zillow. I went to the University of Waterloo in Canada to study computer science and statistics there, and after a few internships, I ended up at Zillow. I've been here for about a year and a few months now, and I've been really enjoying it. 

In the first portion of my time, I was a classical data engineer, just building pipelines. But about eight months into it, I kind of pivoted a little bit to building tools for other data engineers, and that's what I'm doing now.

What got you interested in mixing computer science and statistics?

The way that I ended up in computer science is kind of an accident. I was first studying actuarial science to be an actuary, which is very statistics heavy. I did some internships in that field. And in both of those internships, I wasn't that interested in the actual actuarial work; I was more interested in asking, how could I automate this? How can I automate that? How can I make this report easier to get? That started pointing me in the direction of pursuing something closer to computer science rather than statistics, but I had already kind of finished half of my statistics degrees, so I decided just to go through with it. 

Tell me more about the internship that got you into data engineering.

I was working in a fullstack position. I basically tried to make every internship as different from each other as possible, just so that I could really figure out what I like, and so the last one had me working on a few different projects. The company that I was working for had an agreement with this other company to share their data. At the time, there was not a great way to share that data, to process that data, or to ingest that data. And that was one of my projects — it just seemed really cool that you could do that. 

I don’t really like working on the front end because you put all this work into making a button look a certain way or making a signup form look a certain way. What I like is working on a larger system that has more responsibilities, and more complexities. That is what I got a taste of when I worked on that project.

What are your day-to-day duties like?

When I was a classical data engineer, I was basically building pipelines. We have product teams like that work on the [Zillow] app or that work on the website or that work on other tools that customers use, and then we have to somehow get all that data, and ingest it and process it and serve it to our data scientists and our machine learning engineers in a way that's easy to access and easy to use, and hopefully get it a little bit cleaned up. 

Building pipelines generate reports, stuff like that. A problem that we're facing all the time is that the data quality is not fantastic. Sometimes we got fields in the data that were really not supposed to be there, or data that didn't look like we expected it to, and that would get processed and then the data scientists using that data would say hey, why are my reports coming out faulty? 

We, as the data engineers, are kind of in the middle, since we’re not generating the data. So we had to go back to the product teams, and ask them, where did this data come from? Why does it look like this is supposed to look like this? Are you changing something? And so that was kind of a very time-consuming process. And if you don't catch it, it could generate bad results. 

So I made a little prototype focusing on data quality. We know what the data is supposed to look like, we know what people need from it. I asked myself — is there any way that we could automate this to see if it is quality data or not? So now [instead of a manual check], before we process as part of the pipeline, if something is off, we just get alerted right away [by the automated system]. 

That was a really fun project for me, and that turned into the work that I'm doing right now on my new team. So I pivoted to a team called Data Governance, where they're focusing on data visibility, insights, privacy, and quality. 

Now my day-to-day work now is taking that tool that I built just for my team, and analyzing it to make it accessible and ready for use by the other data engineering teams. And at this point, we're adding more features into it and expanding to different areas so that teams outside of data engineering could also check their data quality before using it.

What skills have you come to rely on most as a data engineer?

Programming, debugging, and Googling things. A lot of these technologies were not familiar to me when I first started. And even if they are familiar to you, you won't know every error message until you encounter them. It's a lot of investigating to figure out if something's not working, and you have to be really curious and willing to persevere. 

Because we're working with different teams and different systems that they're using, they also have to be able to design for how the product teams might be using different technologies and might need to test their data quality, and their build tests versus the streaming teams might need to test their data quality when it's actually being streamed. You have to be very able to go down these paths that you're not familiar with, and investigate and learn everything that you need to know before you build the solution for them. 

As for soft skills, you have to be comfortable with talking to customers, getting their needs, getting their feedback, getting their pain points, and explaining to them why this is something that they should use.

How do you keep up in an environment like that? 

It's something most programmers and most programming fields have to do unless you're working at a bank and they're using 20-year-old technology. 

You have to be curious. I think that's the most important thing, otherwise, it can get a little bit overwhelming, like, “Oh, I just finished this and I have to learn this other thing and this other thing,” like a constant hamster wheel. But if you think of it as, I've tackled this, and now there's this new thing that I could focus on. So I don't get bored. I think it's more of an opportunity to learn new things and to grow skills. 

What are your favorite aspects of the job?

My favorite thing is just knowing that we're building something that helps all these teams and helps all these people drive business insights and make data-driven decisions. I think everyone who works with data knows the saying, Garbage in, garbage out.” And so we're helping them figure out if there's garbage going in. 

My favorite thing is when people come back and they tell me, “Oh, I use this tool, and actually it helps me catch this error that otherwise, we wouldn't have caught and would be propagated down to three other teams and then get back to us.”

Why do you think data engineers are so important, especially with the rising issue of dark data?

I've definitely seen data hoarding, with people being hesitant to throw away data that they haven't even looked at for years. You shouldn't do that. You should be making data-driven decisions, and every decision that you make should have data that backs it up.

We bridge the gap between data collection and data analysis and make it easy. We make it more accessible for the machine learning engineers and for the data scientists to be able to get those insights that they need. 

What’s the gender balance been like during your education and at your workplaces?

At this point, Waterloo is at 18% women in their computer science department. Which is not in the best fit. I used to think that maybe it's not so important, I'm the only girl, so what, it shouldn't matter. 

But it really does matter. I’ve worked at a place — not Zillow, mind you —  where I was the only woman on a team of about 40 people and that definitely made me stand out, you feel you're being scrutinized a little bit more, you have to prove yourself a little bit more. Maybe some of the jokes that they make are not what they would be making if there was a better balance.

And so after having that experience, I started to actually seek out companies and in my interviews I would ask, “Do you have women in leadership positions? What’s the gender ratio?” I think there's work to be done, but it's improving. And I definitely think that it's some improvements to be made.

How do you think women in tech can advocate for themselves?

I think it was really beneficial for me to seek out other women who are working in tech so that you can share experiences, share advice, and build each other up. In the beginning, I felt like I was constantly adding extra emojis and adding extra exclamation marks and saying sorry, too much in my emails, and other people weren't doing that. My friends and I talked about it and we found an article about why women feel this way, and how you should approach it and how you should talk about yourself and emulate confidence. 

That's definitely something that I've been very conscious of since that point. So just having people understand and are in the same boat and share your experiences has been really helpful. 

When you’re looking for a job ask those questions when you're interviewing. As to how many women are in leadership positions that come from different backgrounds, what kind of programs do you have to support that? So for example, Zillow has a Women's Impact Network where they do a lot of workshops and they bring in speakers. Knowing that there are programs like that happening and that they're investing in the women at the company, that was a great signal for me. And that's something that people could also ask for as an indicator.

What do you think that workplaces and educational institutions can do to encourage better gender balance?

First of all, you have to make sure that people want to come to that environment, and make sure that it's actually a friendly and open environment to everyone, no matter how they look or where they come from.

You have to show you care about that by putting your time and energy and funds into other programs. This could be sponsoring hackathons for underrepresented groups, hiring interns from programs like ADA [Ada Developers Academy], which is a programming school for women in Seattle. It’s basically about showing that you understand why it's important to have a variety of experiences and perspectives at your workplace and making sure that everyone who comes in with that perspective feels heard and appreciated and understood. 

This interview has been edited for brevity and clarity.