Yangqing Jia created the Caffe framework while a graduate student researcher at UC Berkeley. He later was a member of the Google Brain project and recently joined Facebook, working on various aspects of deep learning research and engineering.
At the Alliance’s February 2016 tutorial on deep learning for computer vision using convolutional neural networks and Caffe, Jia gave a presentation about implementing deep learning in resource-constrained systems such as mobile phones and embedded devices.
Jia's presentation is available for download here (530 KB PDF).