Meet a Machine Learning Engineer: Q&A with Skyler Grammer
To continue our Q&A series, I sat down with Skyler Grammer, machine learning engineer at Calabrio (and rocket scientist), to learn more about how machine learning is becoming the secret weapon in business.
Below, Skyler talks about how data analytics leads to better decision-making than gut feelings and how machines can go from naïve to brilliant over time.
Q: Tell me a little bit about your background.
Skyler: I have a PhD in astrophysics from the University of Minnesota. I became interested in machine learning in grad school and realized that I was much more passionate about machine learning and programming than astrophysics.
Q: What first interested you about data analytics?
Skyler: I don’t think I can pin down a specific event that triggered my interest in the field of data analytics. Early on as a kid, I realized that the outcomes of my decisions made through the assessment of facts were more likely to be positive than if I were to make decisions based on feelings or subjective perceptions. I suppose this realization was my first introduction to the concept of “scientific thought” and part of the reason that I have always been drawn to statistics and math.
Q: Why is data analytics an important field?
Skyler: The simplest reason that the field of data analytics is so important, in my mind, is that it removes the likelihood of important decisions being made based on gut reactions or just guesses. The more you know—actually know—about a situation, the better your decisions are likely to be. Data analytics as a field is entirely centered around understanding situations through facts and using those facts to educate the people who make important decisions.
Q: What’s one misunderstanding someone has had about your role?
Skyler: The most pervasive misunderstanding is that machine learning can magically predict anything from nothing. Additionally, the distinction between the roles of a machine learning engineer and a data scientist seems to be blurred. The goal of a machine learning engineer is to create the model that describes some process with the least amount of error. The goal is not to optimize for model interpretability, but to make the most accurate and general model.
Q: What’s your favorite part of a new project?
Skyler: In general, the most enjoyable part of any project is watching the evolution in model predictions or results over time. Often, early on the models are naïve, bizarre or hilariously wrong, particularly when using text data. As we see improvements in the model, the process of optimization becomes addicting.
Q: What does a business stand to gain from investing in analytics? What could they lose out on if they didn’t?
Skyler: In business, the goal of data analytics is to use the data that businesses routinely collect to make better decisions. Using data to better understand and predict outcomes is, or at least should be, a pretty enticing thing. Without data analytics, potentially incorrect assumptions, not-yet validated business practices or gut reactions are part of the decision-making process.
Q: What’s one tip you might give a business about how to approach advanced analytics?
Skyler: In my experience, the biggest mistake businesses make when trying to become involved in machine learning or data science is not having a clear idea of how they wish to use machine learning or data science. Once they know how they want to use machine learning or data science, then it is imperative to determine whether the data they currently collect supports their use cases or if they will have the ability to collect the necessary data.
Q: When you aren’t working with data, what do you like to do in your free time?
Skyler: When I’m not at work, I’m usually rock climbing or working out.