AI Research Engineer: LLM & Multi-Modal Pre-Training

Winclap; United States; Remoto; USD 180.000 - 260.000; Tempo pieno

Descrizione del lavoro

Tether is seeking an ambitious AI model researcher to drive architecture development for LLM and multi-modal systems. You will push the boundaries of pre-training, alignment, and data curation to deliver scalable, high-performance models.

You'll work with PyTorch and Hugging Face in a distributed training environment, optimizing token efficiency and cross-modal understanding while collaborating with global teams to advance the field and shape the future of digital intelligence.

Competenze

• Eligibility requires a degree in Computer Science or related field.
• PhD in NLP, ML or related field with strong publications preferred.
• Experience with large-scale LLM or multi-modal pre-training on distributed GPUs.

Mansioni

• Lead large-scale pre-training for LLMs and multi-modal models on distributed GPUs.
• Design and prototype cross-modal architectures and alignment layers.
• Source and curate massive-scale text and multi-modal datasets.
• Plan and run experiments, analyze results, and refine training methods.
• Identify bottlenecks and optimize model efficiency and stability.
• Contribute to distributed training systems for scalable performance.

Conoscenze

LLM & Multi-Modal expertise

Pre-training optimization

Distributed training

Data curation pipelines

Experimental design

Model deployment

Formazione

Bachelor's in Computer Science or related

PhD in NLP/ML preferred

Strumenti

PyTorch

Hugging Face

Descrizione del lavoro

Tether is seeking an ambitious AI model researcher to drive architecture development for LLM and multi-modal systems. You will push the boundaries of pre-training, alignment, and data curation to deliver scalable, high-performance models.

You'll work with PyTorch and Hugging Face in a distributed training environment, optimizing token efficiency and cross-modal understanding while collaborating with global teams to advance the field and shape the future of digital intelligence.