Get to know Revionics Machine Learning Scientist, Charles Lindsey.
We caught up with Charles to talk about what’s coming for AI, how he got started in the industry, and his best advice for anyone wanting to learn more about data science.
What led you to Revionics, and what do you enjoy most about your work?
Using Bayesian methods in the scenario of pricing optimization sounded fascinating. I wanted to learn more about that process.
I enjoy developing tools that can be used on real data and make a substantial difference to companies and the way they conduct their business. Our solutions provide our clients with helpful insights and strategies so they can make informed decisions. To have that kind of impact is exciting.
How did you get started in data science?
I was always interested in computers growing up, and in college I got more into mathematics as well. After graduation, I discovered statistics, where mathematics and computer science were used to get useful information from real data. This process really intrigued me. I enjoyed the challenge of analyzing data with complicated models and ending up with useful answers. So that’s why I gravitated toward data science, and it only gets more exciting as the techniques and technology get more sophisticated and advanced.
What’s an exciting project you are working on right now?
I am working on a project to produce tools that easily give clients a quantified “degree of belief” in their estimates and forecasts. This includes measures of precision, statistical significance, and diagnostics that validate the application of models.
What impact do you see data science has on the retail space?
In grad school we learned how basic statistics was used in the mid-twentieth century to drastically improve manufacturing operations. I think the same thing is happening now in retail. We’re using data science to provide more optimized solutions for retail companies.
Where do you see the future of AI going?
I don’t think the world is going to be completely overrun by robots or anything like that, but I do think AI is going to become ubiquitous. Everything is going to have the possibility to be optimized specifically for us, from your car, to your thermostat, and certainly our phones. I think we’ll have the capabilities to make data driven decisions about every single facet of our lives, large and small. However, there are some hard facts about modeling and mathematics that we aren’t going to be able to fully overcome, so like I said, no conquering robots.
What are some newer AI trends you’ve seen that you’re really excited about?
Well, I mentioned Bayesian methods earlier, but also, there are many newer machine learning algorithms that are not fully studied. We have started to loop back around on some of those and understand them a lot better. Some of the methods were like a black box -you put the data in and get an answer out, but that’s all you see. You couldn’t know how precise an answer you had. You didn’t learn anything about what was probable in the population you were studying, or determine statistical significance. Now we are starting to get the full picture of some of these methods, and we can determine precision and statistical significance. So users get more informative answers when they use these.
What do you do to continue your learning in this space?
There are a lot of good blogs and articles out there, and even reviewing journals and old textbooks can help give you new ideas. Also, I continue to learn just going about my daily work. As I encounter problems and have to find solutions, I’m challenged to come up with new ideas and expand my knowledge around data science.
What advice would you give people who are trying to get into data science now?
Just looking and learning on your own can be good and can get you excited about data science, but you need to get a really good understanding of the basics. That’s really done best through formal education and peer-reviewed resources. Blogs are great for ideas, but you’ve got to read an actual textbook to really get the knowledge. You need a deep understanding of the theory underneath it all.
You were a judge for the Texas A&M Institute of Data Science (TAMIDS) Data Science Competition. Can you tell us a little bit about that?
Overall I was very impressed with the quality of their presentations. Even the undergrads were fairly impressive, with the level of theory and models they used to get their numbers. But beyond that, they also realized how important visualization and presentation are, and were able to explain the results in a succinct way. It was encouraging to see students understand the high-level stuff, but also understand the bottom line, like, what’s the recommendation, and how do I convey all this information? Because in our job, that’s what you need to be able to show the client at the end of the day. I’ve been a judge for a number of years, and I enjoy giving these students a kind of warm up to the industry. It’s a great opportunity and exposure for them. Each year it gets better and the level of competition is higher.
And just for fun, what are you currently bingeing on Netflix?
I’ve recently been watching Kingdom on Netflix. It’s set around 1600 in Korea and it’s like Vikings and Walking Dead basically, and Game of Thrones, too. There’s a lot of interesting stuff going on in the show.