We're back with a new edition of our Spotlight Series featuring data scientist, Ryan Valenza. He comes to us by way of NASA and The Allen Institute for Brain Science, bringing a background in applied physics and math that's deep as it is broad. Keeping him challenged here is the daily pursuit of faster, smarter machine learning models that make brands more efficient in their e-commerce and digital advertising channels. Here's what Ryan had to say:
Tell us about your life before Stackline.
I have an eclectic background. I’m from Baltimore and went to university there to focus on math and physics. To support my education, I worked in technology for a cyber-security company. It was a great experience, but after I graduated, I decided to go onto graduate school to pursue a PhD in Physics at the University of Washington. I got to do a lot of x-ray radiation research and emphasized materials science, so it didn’t closely relate to the e-commerce-focused world I’m in now, but it was definitely fun.
With a background in material sciences and physics, and Blue Origin and Boeing in your backyard, did you ever consider moving into the aerospace industry?
I actually did. I worked at NASA in Houston for a summer in the advanced propulsion lab. That was a very hot summer, and a very exciting one. I have always valued the opportunity to wear many hats and do many different things, and that role certainly provided me with the opportunity to explore out beyond the fringes of my experience up to that point.
You no doubt had the opportunity to return to NASA after you graduated. What did you choose instead?
I took a role at The Allen Institute for Brain Science. I’ve always been split between the desire to pursue pure scientific research and the desire to pursue more of an applied programming and applied math route. After my graduate program, I wanted to do something that was a little bit of both. At NASA, I would have focused more on research – and while there would certainly be an application to that research, it would have been years away. I also considered going to a more conventional tech company, but that wouldn’t have had much of a research emphasis. That’s where a data science role started to make a lot of sense; it straddles the theoretical and the applied.
Can you share more about the work you did at The Allen Institute of Brain Science and some of the triumphs and challenges that led you to consider a role at Stackline?
After graduating from UW, I worked briefly at a startup that gave me a taste of what that environment could be like. I left that opportunity because I was working remotely and wanted to get back to the energy of a shared workspace. I joined the Institute as a data analyst working for the optical physiology team. The Institute as a whole is really working on one experiment: studying individual neurons in the brain to discover how they behave in response to various stimuli. The goal is to map the visual cortex.
Fascinating as that was, I’ve always been attracted to things I’ve never done. I’ve never worked in e-commerce, and I know it’s a blossoming field, so I wanted to get experience with it. I also wanted to work explicitly as a data scientist building machine learning models, and that was something I hadn’t really been getting previously in more pure research or data engineering roles.
I realize my background is fairly unconventional, but I do see an underlying theme in my experiences: through all of them, I’ve been working with computers at the convergence of advanced mathematics and technology.
What attracts you to machine learning projects, and are you getting experience out on the forefront of machine learning technology here like you imagined?
Yes, absolutely. I really enjoy the feeling of discovery that you get when you build an ML model and view the results for the first time. For some reason, even though I understand how a model is working, I'm always pleasantly surprised when it gives me a correct result. We were recently training a computer how to read text from an image, and it’s all driven mathematically, but there’s still a little bit of magic when you see it working with such precision.
What does it take to thrive as a data scientist at Stackline -- both the innate characteristics and key learned skills we would look for?
The most important skill is the ability to pick up new skills quickly, but you do need an extensive mathematics background to work in this field. I also had a lot of experience with Python and C++ as part of my academic background and the work experience I got as I pursued my degree, and that's proved very useful.
This may seem a little counterintuitive since we’re the quietest team, but the ability to communicate complex ideas clearly is extremely important. It’s critical to be able to explain your model to your peers for validation and feedback.
Startups can be viewed as a risk. You talked about prior startup experience priming you for life at Stackline. How would you coach someone through a comparison of Stackline and an institution or organization of Amazon or Microsoft’s heft?
There are much greater rewards here. You get to be a leader and own your projects. You are definitely not a cog in the machine. You’re building the machine.
How is Stackline pushing the industry forward, and how would you describe the biggest opportunities as you look out on a 6-month or 12-month horizon?
I think we are one of the few companies applying machine learning principles to e-commerce and using truly state-of-the-art data science tools in our work. I think many other companies are using rule-based models, whereas we’re really taking advantage of the full suite of machine learning tools. As the industry progresses, I see us continuing to be leaders in that regard, and in advanced automation and multi-channel analytics.
You’ve had an incredible array of experiences thus far in your academic and professional career. Who has inspired you in the work that you do?
I am really motivated by the opportunity to do new and different things, and I have had some fantastic managers along the way who have had the same outlook and pushed me to be brave in these pursuits. When I worked at the cyber-security company during college, my manager was a sysadmin with a Master’s in biology. We talked a lot about how much you can learn when you change fields and bring unique experience to the problems of other disciplines. My manager at NASA was also very influential to me; she’s an extremely intelligent human all-around and helped me build new functional areas of expertise and learn how to adapt my skills to new challenges.
Tell me about the team here. I know some young candidates are wary of joining a team that may be in the early stages of building out the rigorous processes you might find at a larger organization. Do you feel like there’s a good balance of process and flexibility?
When I started, we were in the early stages of developing processes to keep our work aligned and to operate on shared infrastructure. Now, we have many people on the engineering team building out a standardized infrastructure, so it’s easier to spin up new models or build a new software tool. You still own your project, but in terms of deploying that project, there is a solid infrastructure in place.
Your team has an amazing work ethic. How do you seek and maintain balance in your life?
You can definitely achieve great work-life balance here. There aren’t strict work requirements beyond completing your projects, and there’s flexibility in how and when that work gets done.
How would you encourage a candidate to prepare for an interview with you and for a role at Stackline on the Data Science team?
Definitely brush up on machine learning fundamentals. But also in the interview, try to demonstrate that you’ve had the opportunity to pick up many different skills. We definitely value a lot of diversity of thought here.