The Colorado Data Science Team represents CU Boulder in machine learning competitions, and is an incredible opportunity to get hands-on experience applying modern machine learning tools to real problems with real data. We meet weekly on Tuesdays 5-6PM in ECCR 1B40.

Subscribe to the Mailing List

Join the Team: Signup Form


Weekly Meeting Structure

  • Practicing academic or industry data scientists give 10-15 minute talks
  • Students give 5 minute tutorials and presentations
  • Team discusses approaches to current competitions
Learn more about giving a talk or sponsoring a recruiting competition

Competitions

Previous Next

Recent News

Veljko Krunic of Krunic Consulting presented on how best to develop machine learning technology to solve business problems. Team members Sachin Muralidhara and Akash Devgun presented on the Pandas data processing library and ensemble models, respectively.
Team members Sarah Withee and Henrik Larsen presented on problems they've been working on, the DARPA signal processing challenge, and the MNIST digit recognition task, respectively. We also discussed ways of tackling the Oracle Data Cloud competition
Weekly Meeting 11/08/2016
Caleb Phillips, Data Scientist at National Renewable Energy Laboratories(NREL) spoke to us about a few projects he's worked on at NREL, involving big-data and geospatial analysis.
Weekly Meeting 11/1/2016
Alvin Grissom, a final year PhD student at CU Boulder, talked about challenges in machine translation from Japanese or German to English such as verb prediction. Ling Liu also presented a Jupyter notebook on using the deep learning library Torch
Weekly Meeting 10/25/2016
Adam Bloniarz from Google Boulder shared some computational neuroscience research findings of his, and team member Daniel Korytina presented on AdaBoost, a popular and well-performing general purpose learning algorithm.
Team member Nehal Kamat gave a short tutorial on how to run a simple model on our Oracle data with Vowpal Wabbit, and Team Captain Pedro Rodriguez showed some Apache Spark code useful in dealing with large amounts of those data.