Description
ZYPHRA is hiring a Research Engineer - Language Model Pre-Training to develop end-to-end language model pretraining systems. The role focuses on large-scale training runs, model parallelization, pretraining-stack performance optimization, dataset collection and evaluation, and architecture and methodology research. Candidates should have strong machine learning and engineering skills, experience with large-scale GPU clusters and distributed training pipelines, high proficiency with PyTorch and Python, and a postgraduate degree in a scientific subject.
