AI Research Engineer (Pre-training - LLM & Multi-Modal)
Core
Drive innovation in architecture development for cutting-edge Large Language Models (LLM) and Multi-Modal systems, focusing on pre-training optimization, data curation, and alignment.
Role type
Senior IC AI Research Engineer (Pre-training - LLM & Multi-Modal)
Builds
Foundational pre-trained models for Tether's AI and peer-to-peer technology initiatives (e.g., KEET app).
Domain
Fintech / AI Research / Large-Scale Distributed Systems
Deliverable
production ML models
Required skills
Large-scale LLM and Multi-Modal pre-training, Transformer architecture design, Distributed training systems, Data curation and pipeline engineering, PyTorch, Hugging Face, NVIDIA GPU cluster management, Cross-modal alignment techniques, Experimental research and debugging.
Preferred skills
PhD in NLP/Machine Learning, Publications in top-tier AI conferences (A*), Experience with non-transformer modifications.
Technologies
PyTorch, Hugging Face, NVIDIA GPUs, Distributed training frameworks.
Responsibilities
Conduct foundational pre-training for LLMs and Multi-Modal models on large distributed servers; Design, prototype, and scale innovative architectures and tokenizers; Source, filter, and curate massive-scale textual and multi-modal datasets; Execute experiments to refine training methodologies; Investigate and resolve bottlenecks in model efficiency and alignment; Contribute to the advancement of distributed training systems.
Seniority
Senior, hands-on IC