October 3, 2013
Last Wednesday, Vendavo headquarters hosted the Silicon Valley Big Data Science MeetUp, and I felt more excited than ever about the new algorithms and infrastructure demonstrated by OxData (view their GitHub here). MeetUp events like these provide the technical community with the chance to exchange ideas, discuss new advances, and generally stay abreast of the latest trends.
The meeting began with casual conversations over pizza, beer, and soft drinks, where data scientists young and old shared current and past experiences. The crowd was filled with a wide range of backgrounds, from engineers that had worked on 1000 node DB2 clusters, to weather researchers working on super computers to conduct abrupt climate change research. The evening highlighted the two classes of data scientists:
– The ones who are doing pure data research, using tools like H2O to perform their large data explorations
– The ones whose research demands new algorithms and/or new implementations
And, OxData brought the goods for both groups.
They started off with a presentation (see slides here and here) on their new implementation of Gradient Boosting. It was a well-prepared talk, describing the algo’s purpose, where it is best applied, and why. Then, for the hackers, they went below the hard deck and described the programming models within their H2O computation engine for implementing Distributed Trees and Map Reduce based algos.
It was great fun to raffle off a few books, the least of which wasn’t the Bad Data book. I love moral judgments on data!
I think the success of the night can be judged by how long people continued to hang out and talk after we were done, stoked on the intellectual stimulation of the evening. In fact, we had to start picking up chairs to get people to leave. I look forward to the next event at Vendavo and would love to hear suggestions for new topics!
If you’re interested in attending future Vendavo Tech Events, sign up here!