I'm currently a research scientist at a small startup, Intellisis, in San Diego. I spend most of my time developing neural networks for signal processing. Now some people think neural networks are completely retro but there is a lot of new work that goes well beyond simple sigmoidal networks trained with back prop for example deep belief networks, Google's pregel, sparse coding. There's also a lot of hardware being developed to implement neural networks (e.g Darpa's SyNAPSE project,SpiNNaker, and Neurogrid.

I spend most of my time these days dreaming up new ways to use neural networks and then building those neural networks using C++, CUDA, and python.

Before moving to San Diego I was doctoral student in Georgia Tech's Bioengineering program. My research addressed the problem of making optimal decisions in an uncertain world; an area of research variously referred to as active learning, reinforcement learning, and sequential optimal experimental design. I studied this problem in the context of optimizing neurophysiology experiments to collect data efficiently. For more information please follow the link to my research. Most of my research was in collaboration with Liam Paninski.

I'm a former member of Georgia Tech' Neurolab. The Neurolab is a collection of faculty and students working on related problems in neuroengineering. The members of the Neurolab share resources and space, which makes its collaborative setup rather unique in academia.

I was also one of the students in the Neurolab's IGERTfor Hybrid Neural Microsystems (HNM). An IGERT is a training program funded by the NSF to promote specific areas of research. The Neurolab's IGERT is focused on interfacing neural systems with MEMs devices to better understand neural processing and neural injury.

My research was supported by the Department of Energy's Computational Science Graduate Fellowship (CSGF). The CSGF was created to support students using high performance computing (HPC) to solve problems in engineering and science. In my research, I use (HPC) to optimize decision making by enumerating the possible outcomes of a decision and their associated value. A major computational challenge of my research is solving these problems in "real-time" so that the solution can be used to make actual decisions.