Remote AI Research Engineer: Model Compression & Multimodal

Tether Ltd.; Craigavon; Hybrid; GBP 90,000 - 150,000; Full time

Job summary

Tether Ltd. is seeking an experienced AI research scientist to drive model compression for multimodal systems, including large language models and vision-language models.

You will reduce model footprint and compute costs while preserving accuracy, applying quantization, knowledge distillation, and pruning to complex architectures that integrate text, images, and audio. You will build robust compression pipelines, establish performance and fidelity metrics, and address bottlenecks in production

Qualifications

• PhD in NLP, ML or related field preferred.
• Experience with PyTorch or equivalent frameworks.
• Hands-on in QAT and PTQ quantization.
• Knowledge distillation for compressing large models.
• Model pruning experience.
• Strong understanding of neural networks and transformers.
• Familiarity with C++ is a plus.

Responsibilities

• Apply low-bit quantization to reduce model size and inference latency.
• Leverage knowledge distillation to transfer capabilities from larger to smaller models.
• Implement pruning techniques to remove redundant parameters and attention heads.
• Analyze trade-offs between model efficiency and accuracy across methods.
• Develop robust compression pipelines and metrics for production inference.
• Publish findings in top-tier conferences to advance the field.

Skills

Quantization

Knowledge distillation

Model pruning

Multimodal AI

Edge deployment

Education

PhD in NLP / ML / related field preferred

Bachelor's in Computer Science or related field

Tools

PyTorch

C++

Job description

Tether Ltd. is seeking an experienced AI research scientist to drive model compression for multimodal systems, including large language models and vision-language models.

You will reduce model footprint and compute costs while preserving accuracy, applying quantization, knowledge distillation, and pruning to complex architectures that integrate text, images, and audio. You will build robust compression pipelines, establish performance and fidelity metrics, and address bottlenecks in production