August 6, 2020
Colin Brochtrup recently celebrated his one-year anniversary as a data scientist on the GK team. I sat down with him (virtually) to see how well he’s managed over this tumultuous past year. Colin discusses how he keeps busy in his spare time, his love of building and making things both inside and outside of work, and how GK’s culture has helped him to challenge the idea that the most important thing is “being right.”
Hey Colin, thanks for taking the time to talk with me. Can you tell me a little about yourself and your interests outside of work?
It may surprise you, but I enjoy NLP (Natural Language Processing) so much that I’ve used it in several personal projects outside of work. My current project is an NLP powered vocabulary game. I’m using it to expand my vocabulary, and it gives me the opportunity to play with state of the art systems.
When I get sick of looking at computer screens I improvise piano pieces, whittle wood, or swing dance (for now that’s alone). Recently I delved into shrubs– which are cocktail mixers, not bushes– made from fruit, herbs, and vinegar. I’ll try pretty much any hobby that makes something!
What brought you to GK, and why is it a good fit for you?
My Master’s degree was focused on speech processing, which was perfectly suited for GK’s analysis of voice data. But of course, GK’s culture of relentless improvement was what sold me. After I submitted my technical assessment I thought I would never hear about it again. But to my surprise, during the on-site interview we methodically reviewed my work together, and we fixed all of the issues with it. That candid and constructive feedback convinced me that I would be challenged to continually improve at GK.
How did you get into the field of data science?
I first became interested in making decisions with data and math in a high school statistics course. For a while, I meandered between mathematics, physics, and electrical engineering. Then I got laid off from a job, and that gave me the time and opportunity to research speech recognition, speaker identification, and audio noise identification. Doing that research gave me experience with a bevy of deep learning techniques to qualify me for a data science position at GK, and the rest is history!
What do you most enjoy working on?
I love the final stretch of implementing a new system, when you’re hooking everything up and you finally get to see your model’s output on the front end. It’s rewarding to watch a dozen small tasks, all working together to form a system that unlocks value for customers.
What’s the most important aspect of your role?
For me, it’s avoiding “analysis paralysis.” I tend to overthink possible solutions to complex problems, when it’s better to develop a few candidate solutions and create simple versions in order to get back information on their efficacy. When I’ve asked senior data scientists for advice, a lot of their feedback has boiled down to “Just try it.” At first, it was tough to override the voice in my head telling me “But what if this is the wrong method?!” A year at GK has helped tremendously to build my confidence in solving problems, and to produce minimal systems that can be iterated on.
Do you have a favorite mantra or quote?
“I could be wrong” has served me well professionally and personally. It keeps my mind open when entering any discussion, and has helped me learn from everyone with whom I interact. At my previous positions, I saw how deleterious it was to be focused on “being right” (instead of finding the best solution). I appreciate how every member of the GK team discusses designs or solutions to existing problems, and I think it’s a major reason for our rapid and fastidious development.
Who do you most admire at GK and why?
Matthew Goldey, one of our Senior Data Scientists, who definitely champions GK’s cultural principles. Every day I see his compassionate candor, attention to detail, his problem-solving skills and his professionalism, and I think, “Man, I hope I’m that good someday.” Working closely with him has taught me a lot about being a good data scientist, and even more about being a great co-worker and teammate.