I am Yiğit Demirağ, an EEE M.Sc. student at Nanotechnology Research Center (NANOTAM) at Bilkent University, working with Prof. Ekmel Özbay. Currently, I am a research intern at Memory Division of Samsung Electronics in Korea. Previously, I was at EPFL, Google and CERN. My focus is to build a perfect accelerator chip for Neural Networks. In my research, I design brain-inspired memory devices using Phase Change Materials (PCRAM) and study multilayer supervised learning on chip. In my spare time, I experiment with neural network architectures and read about in Deep Reinforcement Learning. I usually code in C/C++, Python and MATLAB. To learn more about me, you can read my tweets or drop me a line.

You can read my blog here.






2015-now
Bilkent University: Master's Degree
Designing comprehensive PCRAM simulation.
Advisor: Prof. Ekmel Özbay.
Summer 2017
École Polytechnique Fédérale de Lausanne (EPFL)
Multilayer Supervised Learning
in Neuromorphic hardware.
Advisor: Prof. Yusuf Leblebici.
Summer 2015
Google Summer of Code
Worked on ARM Neon extension of Agner FOG's VCL.
Summer 2014
CERN
Vectorization of Counter Based
Random Number Generators.
2013-2015
Middle East Technical University
Hardware design of particle (Muon) detector.
Advisor: Prof. Bilge Demirköz.
Summer 2011
Bilkent University Computational Electromagnetics Research Center
Acceleration of world record holder
MLFMA algorithm by threading.
2010-2015
Bilkent University: Bachelor's Degree
Major in Electrical and Electronics Engineering.