Research Scientist, Distributed Pretraining

Pluralis Research in Australia is advancing Protocol Learning to train frontier-scale models using distributed training across consumer hardware. You will design and optimize model-parallel training, ensure robustness over networks, and push the boundaries of how large models run on off-the-shelf devices. As a Research Scientist, you will ship systems, contribute research code, and collaborate across time zones with teams in Australia and North America to make scalable, decentralized AI trainin…