Xilinx has intern positions available for summer 2019 in the area of machine learning.

The position is in San Jose.  Please send your resume/CV to Prof. Fred Harris at fjharris@eng.ucsd.edu

1. Scheduling beamforming in massive MIMO 5G networks

This project investigates of ML techniques for beam scheduling in a massive MIMO network. This https://spectrum.ieee.org/tech-talk/telecom/wireless/3-ways-nokia-is-using-machine-learning-in-5g-networks short description from the Nokia CTO provides a brief overview of the problem. This project would involve coding a simulation environment (probably in matlab, but there are other options) for a MIMO system and generating data to be used in training a network for beam steering. Of course the ML network would also have to be researched and coded in TF, trained and the key performance metrics determined. This project would be best suited for a graduate student that has a strong ground in wireless signal processing and that has also done some ML classes and has some experience working with ML networks.

1. Implementing RNN networks on a Massively Parallel VLIW DSP Array

In 2018 Xilinx announced a computing platform called ‘Versal’. The Versal device contains a large multicore array of vector signal processing engines that are optimized for several domains including machine learning. These processors are programming in C/C++. This project would involve hands-on coding, in C/C++, of LSTM-type networks for the Versal multicore array. This project is more suited to a graduate student that has an eye to implementation and has an appreciation for parallelizing algorithms for parallel hardware. Good knowledge of C/C++ coding would also be required. SO this project is more for a student with an orientation to software development.

1. Ensemble Learning, e.g. Random Forrest, on multicore VLIW DSP Array

In 2018 Xilinx announced a computing platform called ‘Versal’. The Versal device contains a large multicore array of vector signal processing engines that are optimized for several domains including machine learning. These processors are programming in C/C++. This project would involve hands-on coding, in C/C++, of random forrest-type networks for the Versal multicore array. This project is more suited to a graduate student that has an eye to implementation and has an appreciation for parallelizing algorithms for parallel hardware. Good knowledge of C/C++ coding would also be required. SO this project is more for a student with an orientation to software development.