FPGA-based Machine Learning Acceleration
| Company: | CruxML |
| Duration: | 3-months |
| Start Date: | November 2023 |
| Location: | Sydney – Darlington |
| Scholarship: | $20,000 |
About the Company
CruxML provides services and solutions for machine learning that achieve real-time performance on FPGA platforms for real-time intelligent sensing. CruxML’s mission is to work with our clients to adopt real-time machine learning in their applications.
Project Objective
The intern will conduct numerical experiments in the training of quantised deep neural networks to support real-time radio frequency machine learning.
The project is to test newly developed inference techniques for prediction on our projects. To date, we have been using ResNet but there is a vast literature on improved architectures which may be more amenable to efficient FPGA implementation, e.g. RepVGG. The deliverable will be a report detailing the different techniques explored.
Intern Skills
Experience with designing convolutional neural networks or other forms of deep neural networks. Experience with Python and TensorFlow or PyTorch.
Application
Applications must be submitted online by 24 September 2023 and should include additional documents:
- Curriculum Vitae
- Motivation Letter
- Supervisor Support Letter
- Proof of Australian citizenship or permanent residence
(e.g. passport, birth certificate, citizenship or PR certificate)
Our Partner Universities
