Home page
This page is a bit outdated (2 years old), I graduated and am now working at NVIDIA as a Senior Deep Learning Architect, working on efficient algorithms, future hardware architectures and performance optimization for accelerating deep neural networks with a focus on NLP and computer vision. Please see my linkedIn for most up-to-date information.
I did my PhD in Electrical Engineering and Computer Science departments at Stanford University. My main research area was neuromorphic computing with a focus on hardware implementation using emerging memory technologies. Neuromorphic (brain-inspired) computing aims to mimick the real time processing power and energy efficiency of human brain for tasks that are too complex for conventional hardware (e.g., your desktop). It has great promise especially in IoT era, where 'energy efficient real time computing' is precisely what is needed. Using emerging nanoscale memory devices such as RRAM, PCM, CBRAM, etc. as 'analog' memory and storage units (synapses), and by closely integrating them with massively parallel computational units (neurons), I design intelligent adaptive hardware that can asynchronously process data without seperating processor and memory units, hence making them extremely energy efficient. This paper that was presented in IEDM on 12/2015 and this paper that I presented in ISQED on 3/2016 give a good introduction to the field and my research.
During my PhD, I have been very lucky to be collaborating with groups that are experts in circuits, architecture and algorithms, both from Stanford and other schools and research institutes (UCSD, IBM, Penn State, Tshingua, ASU, etc).
Besides my PhD research, I have worked on statistical financial modeling and forecasting. I served as the Financial Officer of Stanford Windsurfing Club and Stanford Turkish Student Association for 2 years.