Noah Randolph graduated from the School of Information with a Master of Information and Data Science in 2018. He currently works at Picarro as a manager and data scientist, where he is building and improving predictive models for production as well as leading data-driven decision making. Prior to the I School, he graduated from the Georgia Institute of Technology with a bachelor’s in international affairs and mechanical engineering.
What have you been up to since completing the program?
I joined my W201 instructor at a startup called Kountable (now KLEAR), which is a place where dealmakers in Africa go for help with trade funding, procurement, and logistics. I was hired to do a broad range of data analytics and machine learning. Early on, I implemented a regressor using AWS Lambda functions to filter out new project submissions that are very likely to fail our process. However, while working on that, investors began to say we needed a more scalable platform, so I switched over to a developer role, leaving data science for a future day while we handle the immediate scalability concern.
How has the program prepared or helped you in your current job?
Luckily I am using lots of what I learned from the program, despite not being a data scientist by title. I took on the task of designing our new operational database schema because we landed on a distributed key-value store, and I recognized what was possible due to what I learned in W205 and W261. We’re developing on a model-view-controller architectural pattern, but the key insight has been to greatly simplify the complicated data manipulation required by our application with Pandas DataFrame tricks that I learned throughout MIDS.
What was your favorite course and why?
My favorite course was W261 simply because I enjoyed the assignments so much. I’m using the same techniques I learned to make a MapReduce or Spark job do what I want in W261 as in my current job, which is in turn why I’m enjoying it so much.
What is the most valuable thing you learned while in the program?
If it wasn’t already obvious from the previous answers, I like the set of skills I gained from the program. I wasn’t sure if I’d ever use what I learned in Data Visualization, but even there I’ve come up with some cool stuff for Kountable. Now all I need is a field experiment in my job and I’d have the whole data science spectrum. Without MIDS, I’m not sure what I would be doing- maybe I would still be sweating like a pig in a chemical plant, but I don’t think I would enjoy it as much.
What one piece of advice would you give to prospective or current MIDS students?
Use the networks. I got two job offers and an internship just from paying attention to what people put out in the community. For prospective students, live classes online (which MIDS does) are the way to go. In person classes require taking the time to go somewhere, and recorded online classes are very boring, requiring you to spend determination capital that could be used elsewhere.
