Summary

This role develops pre-training data pipelines for Cohere's foundation language models, combining machine learning research with software engineering. The position is remote-friendly with location flexibility between EST and EU time zones, and focuses on data quality, curation, mixtures, and training efficiency.

Responsibilities

• Conduct data ablations and test data mixtures
• Develop data modeling techniques for training datasets
• Research and implement data curation methods
• Build data pipelines for language model training
• Collaborate with researchers and engineers on pipelines