Monday, October 08, 2012

On why I'm excited about "big data"

I was in Aarhus recently for a MADALGO workshop on large-scale parallel and distributed models, where I did a sequence of lectures on GPU algorithms. I was briefly interviewed by a university reporter for an article, and did a little video on why I think big data/big iron problems are interesting.

At the risk of embarrassing myself even more than I usually do, here's the video. Note that this was recorded at a time of great crisis across the globe, when all hair styling products had mysteriously disappeared for a few days.

Wednesday, October 03, 2012

We're hiring FIVE (count 'em, FIVE) faculty this year.

We had an incredible hiring season two years ago, making seven offers and hiring seven new faculty. And now we're doing it again !

Our department is looking to hire five new faculty (at least four are at the assistant professor level). I'm particularly excited that we're hiring two people in the general area of big data (following up on our data mining and database hires from two years ago).

One slot is in what I'll call "big data meets big performance": I'll have more to say about this shortly, but the challenges of large data analysis are not just about managing the data, but about managing large numbers of machines to crunch this data (MapReduce is perhaps the most well known example of this). We're looking for people who can "speak extreme data and extreme processors" fluently - these could be on the data/systems management side, or on the analysis side, or the modelling side.

Utah has a strong presence in high performance computing (the Supercomputing confererence is happening in Salt Lake, and Mary Hall is the general chair), and we're one of the few places that has a good understanding of both sides of the large data story (i.e machines and bits).

The second slot is in machine learning, with ties to language. Text (and language) provide one of the best large data sources for scalable machine learning, and we're looking for people interested in the challenges of doing ML at scale, especially when dealing with NLP. If you're that person, you'll be coming into a department that has the entire range of data analysis folks from algorithms to analysis to systems to NLP (with many other faculty that are heavy users of ML technology).

Our plan, once we fill these slots, is to make Utah a one-stop shop for large-scale data analysis and visualization - in addition to the above slots, we're also looking to hire in information visualization to complement our strong viz presence.

In addition to the above slots, we are also hiring in computer security and HCI/user interfaces. While I'm equally excited about these positions, I know much less about the areas :). I will point out that we have a big systems group that covers many aspects of security (language security, verification, and network security) already. We've also had strong demand from students and industry for research in HCI, which will complement our info-viz efforts (and even our data mining work)

For more details on how to apply, see our ad. We'll start reviewing applications after Dec 1. Feel free to email me if you have questions about the slots (but don't send me your application material - send them in directly)

Disclaimer: the above views are my own personal views, and don't represent the views of the department or the hiring subcommittees.

Monday, October 01, 2012

Things a _____ should do at least once...

Without quite realizing it, I managed to create a (tiny) meme in the rarefied circles of TCS/math with my post "Things a TCSer should have done at least once".

Firstly, you should check out the G+ post for even more scathing commentary on my original post.

Next, you should see the followups (in case you haven't already):
Let me know if there are more - I'm still waiting for a quantum computing version.(thanks, Pontiff!)

Disqus for The Geomblog