Nicholas Fraser
Machine Learning Researcher at Xilinx
Dublin, Ireland
Summary
🤩
Rockstar🎓
Top SchoolNicholas Fraser is a Machine Learning Researcher with 11 years' experience specializing in reduced-precision and hardware-aware neural networks, currently driving research at Xilinx Research Labs in Dublin. He combines a PhD in Computer Engineering and first-class electrical engineering training to bridge algorithm design and FPGA/ASIC implementation, from modelling semi-differentiable quantization to synthesizing networks directly into circuits. His work has extended major frameworks (TensorFlow, Torch, Caffe, Darknet) for low-precision training and produced open-source tools and demonstrators, including contributions to the widely used BNN-PYNQ repository for quantized networks on PYNQ hardware. Nicholas also develops robustness techniques for hardware faults, mentors interns and collaborates with academia, bringing both hands-on DSP/embedded experience and a track record of publishing and presenting at conferences. An unusual strength is his ability to move innovations from invention disclosures through prototype FPGA libraries to customer-facing presentations and open-source releases.
10 years of coding experience
1 year of employment as a software developer
The University of Sydney
English
Github Skills (10)
neural-network10
quantization10
machine-learning10
python10
theano10
lasagne10
image-augmentation9
computer-vision8
sys7
embedded7
Programming languages (9)
Github contributions (5)
Xilinx/BNN-PYNQ
Mar 2017 - Jan 2022
Quantized Neural Networks (QNNs) on PYNQ
Role in this project:
ML Engineer
ML EngineerContributions:26 commits, 1 PR, 8 pushes in 4 years 10 months
Contributions summary:Nicholas contributed to the development of quantized neural networks on the PYNQ platform. Their work includes implementing BNN training scripts for MNIST and CIFAR10 datasets and generating binary weight files for hardware implementation. They also added an example for the GTSRB dataset and included image augmentors. Furthermore, the user updated scripts to point to new filenames and updated license information.
quantized-neural-networksneural-networksmachine-learningpynq
nickfraser/optimum-amd
Feb 2024 - Nov 2024
AMD related optimizations for transformer models
Contributions:4 PRs, 92 pushes, 44 branches in 9 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.