Remote AI Research Engineer – LLM & Multimodal Pretraining

Tether Ltd.; Craigavon; Hybrid; GBP 100,000 - 140,000; Full time

Job summary

Tether Ltd. is shaping the future of digital finance with AI-driven models for cross-modal understanding and scalable pre-training. You will lead architecture development for LLMs and multimodal systems, advancing token efficiency and model performance on vast GPU clusters.

Applicants should have a PhD in NLP/ML, hands-on experience with large-scale pre-training, and deep PyTorch/Hugging Face expertise. This is a remote, international role at the forefront of fintech AI research.

Qualifications

• Advanced degree in CS or ML with publications in AI R&D.
• 0
• Hands-on experience with large-scale LLM or multi-modal pre-training on thousands of GPUs.
• Familiarity with distributed training tools and libraries.
• Deep knowledge of transformers and non-transformer modifications to boost efficiency and scalability.
• Strong PyTorch and Hugging Face proficiency for model development and deployment.

Responsibilities

• Large-Scale Pre-Training for LLMs and Multi-Modal models on multi-node GPU clusters.
• Architecture, Alignment Innovation to improve model intelligence and cross-modal understanding.
• Source, filter, and curate massive-scale text and multi-modal datasets; establish data pipelines.
• Execute experiments, analyze results, and refine training methods for token efficiency.
• Identify and resolve bottlenecks in model training and multi-modal alignment stability.
• Contribute to distributed training systems for scalable hardware efficiency.

Skills

Large-Scale Pre-Training

Architecture & Alignment Innovation

Data Strategy

Experimental Research

Optimization & Debugging

System Scalability

PyTorch & Hugging Face

Education

PhD in NLP / Machine Learning

BSc in Computer Science

Tools

Distributed training frameworks

PyTorch

Hugging Face

Job description

Tether Ltd. is shaping the future of digital finance with AI-driven models for cross-modal understanding and scalable pre-training. You will lead architecture development for LLMs and multimodal systems, advancing token efficiency and model performance on vast GPU clusters.

Applicants should have a PhD in NLP/ML, hands-on experience with large-scale pre-training, and deep PyTorch/Hugging Face expertise. This is a remote, international role at the forefront of fintech AI research.