Data Scientist Vincent D. Warmerdam about his Work
Each data science role may be different, but our data scientists provide insight to help those interested in figuring out what a day in the life of a data scientist actually looks like.
"We build real stuff, stuff that works and stuff that makes it into production," Vincent D. Warmerdam, Data Scientist at GoDataDriven, just likes to build systems that solve problems.
Field Notes from a Data Scientist - Vincent D. Warmerdam
Data Scientist (and self-proclaimed “Pokemon Master”) Vincent D. Warmerdam wants to build systems that solve problems. His motto: “Never let your school get in the way of your education.”
One of the projects for GoDataDriven I’m most proud of is the recommender we built for the website of NPO, the Dutch public broadcast corporation, where visitors can watch television programs. The recommendations were working so much better after we did our thing that my grandfather even congratulated me for it. The rest of the country also noticed; moments after implementation, the recommender broke the caching layer due to a shift in popularity - that was rather cool.
There is a lot of bullshit in consultancy. A herd of elite men in suits making powerpoint presentations about the possibilities of data as the core delivery of their service. At GoDataDriven, we build real stuff, stuff that works and stuff that makes it into production.
My current client is Royal FloraHolland in Aalsmeer, where they auction flowers. Our assignment is somewhat generic: we make them better at what they do by using data. We build systems to predict how much traffic enters the site, fluctuations in the prices of particular flowers, and even to deal with the quality of the flower images. Usually, we take two weeks or so to build a proof of concept, and if it appears useful, we bring it into production in a few months.
I don’t see my job as just designing algorithms- I design and build systems that solve actual problems. When I’m gone, the system must continue to work without me and must be understood by the people who have to work with it. Some data scientists like to brag about how complex their algorithms are. Instead, I get more satisfaction from building a system that is as simple as possible.
My motto is, “Never let your school get in the way of your education.” Education has always been an important part of my life. I taught myself how to program. I started my career as a freelance developer. In order to grow my network and give something back to the community, I started to give free courses on data science. Now I teach several courses here at GoDataDriven and through Xebia Academy. I also get to work for major clients like ING, NPO and Royal FloraHolland, where there are a lot of cool projects. Gaining access to these companies and working on projects of this scale would be more difficult as a freelancer.
Another cool part of my work here is that GoDataDriven gives its employees a lot of space for personal projects. With a colleague and friends from the community, I was able to co-found PyData. We organize meetups and a yearly conference for developers and users of open source data tools. We thought that there was a lack of technical events where you could actually learn something, and we were very happy that we were able to work together with NUMFOCUS to bring the conference to Amsterdam. I also initiated a partnership with Rstudio. They’ve always impressed me with what they do, so I was very happy to connect with them. I also share some of the lessons I’ve learned on my blog, koaning.io.
A smart and nice way to learn new stuff is to build something with a data engineer in a new language or new technology and then try to present that project at a conference. Last year my colleague Fokko and I worked on a small hobby project that uses Apache Flink, a streaming technology. We built a model for online video games, to match opponents of the same strength in real time. We showed it to some of the Apache Flink developers, and they were impressed enough that they asked us to submit it to their conference. You can check it here: Flink Forward Berlin - Using ML Algorithms for Online Gaming. Getting asked by the Apache Flink developers to speak at their conference was great, actually being there was a blast.
What I love about my field of work is that when I see that something isn’t okay, I can make an app or tool to solve that problem. The downside of this way of thinking is that it is pretty addictive. At some point, my girlfriend had to tell me that I needed to close the laptop. Now I apply a bit of mental accounting: I try to figure out if the energy going into an activity is greater than the energy I get out of it. If it is, I should not do it. I’ve seen many people in my field getting hit by burnout and I want to prevent that.
I now know better what gives me energy, and I’m getting better at scheduling my attention. I try to do a bit more sports and take the time to turn off the laptop, so I have more time for my friends, family, and my two felines.
Never let your school get in the way of your education.