Here are the details of the Talk. Title: Predictive Models at Scale using Dumbo
Speaker: Nikhil Ketkar Abstract: The data science stack for Python is mature and robust. Libraries like Numpy, SciPy and scikit-learn allow data scientists to build predictive models easily. However, when it comes to making predictions on large volumes remains an operational challenging. Data scientists typically end up using Python just for prototyping models and then implement models in Java so they can leverage Hadoop. With libraries like Dumbo make it possible to build and run machine learning models in Python that can make predictions over very large datasets. The talk will describe the problem and the proposed solution with example code. Speaker Bio: Nikhil is currently the Director of Engineering at Indix (www.indix.com). He received his Ph.D. from Washington State University. Following that he conducted post doctoral research at University of North Carolina at Charotte, which was followed by a brief stint in high-frequency trading at Transmaket in Chicago. More recently he led the data mining team in Guavus, a startup doing big data analytics in the telecom domain. His research interests include machine learning, data mining, and graph theory. On Sun, May 10, 2015 at 4:21 PM, Sharmila Gopirajan < sharmila.gopira...@gmail.com> wrote: > Hi Nikhil, > Looking forward to your talk! > > On a related note, I would like to share some lessons I learned from > ChennaiPy talks. If it helps someone, thats great. As Shrikant has > observed in another reply, the audience will be a diverse group from > different backgrounds. Some of them would be quite familiar with your > field and others might not. It helps if your presentation takes that into > consideration. Based on my experience, if your talk is a more technical > one (than, say, an introductory one), it helps to narrow your presentation > to address just one aspect or feature. That way, even people unfamiliar > with your area can catch up. Also it helps if you try presenting it to > someone first. Trust me, that made a huge difference ;) . If there are > more ideas for you to present, you can always follow it up with more > presentations ;) Probably make it a series. > > Also a link shared by Shrayas has been quite helpful. So I'm just sharing > it again. http://speaking.io/ > > Regards, > Sharmi > > _______________________________________________ > Chennaipy mailing list > Chennaipy@python.org > https://mail.python.org/mailman/listinfo/chennaipy > >
_______________________________________________ Chennaipy mailing list Chennaipy@python.org https://mail.python.org/mailman/listinfo/chennaipy
