Soren Macbeth, Chief Scientist/Data Hacker at Yieldbot and former co-founder of Stocktwits, joined us for the NYC Data Science Meetup last month. Unfortunately, we had a projector issue that prevented us from showing slides, but Soren rolled with it and did the entire talk from memory.
In the video, Soren discusses how Yieldbot operationalized their data science efforts using tools like ElephantDB and Storm. He also touches on his personal philosophy when it comes to research versus production work (and how Clojure fits into that) and some of the challenges Yieldbot has run into when operationalizing machine learning systems.
Here’s a quick summary of the video:
- Background (0:17)
- Building Yieldbot’s data systems from scratch (4:06)
- Why he chose Cascalog (6:30)
- ElephantDB discussion (9:07)
- Yieldbot’s use of Storm (10:57) (also more in Q&A, especially at 20:56-23:40)
Soren’s talk itself goes to 15:40, but the Q&A the follows is full of awesome information. Thanks again to Soren for sharing his experience with us.