Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will conduct novel research on decentralized training challenges and aim to publish at leading ML conferences. You will select a foundational scaling problem, investigate it during a fixed-term PhD internship, and use large-scale training infrastructure to develop your results.
Requirements
- Current PhD candidate status
- At least one publication in NeurIPS, ICML, ICLR, or another top-tier ML venue
- Research focus in a technical area relevant to frontier models
- Strong theoretical understanding of deep learning and distributed systems
- PyTorch proficiency
- Experience with large-scale training infrastructure
- Professional-level written and spoken English
Responsibilities
- Conduct novel research in Protocol Learning
- Publish research in tier-1 ML conferences
- Select and answer a scaling problem affecting Protocol Learning