Research Scientist, Post-Training
Research Scientist, Post-Training
Research Scientist, Post-Training
DatologyAI
DatologyAI is seeking a Research Scientist to lead post-training data curation for foundation models. The role involves designing algorithms for improving datasets and optimizing data across pre-training and post-training stages. The company focuses on enhancing model training efficiency and quality through smarter data selection, backed by significant funding and a team of experts in deep learning.
Qualification
- Strong scientific judgment
- Fluency with deep learning literature
- Experience in data curation and model training
- Ability to work autonomously and collaboratively
- Drive to translate research ideas into real-world applications
Responsibility
- Conduct research on algorithmic curation of post-training data
- Generate and refine preference and instruction-following datasets
- Curate capability- and domain-specific data for post-training
- Unify pre-training and post-training data curation
- Transform complex research literature into actionable improvements
