Remote AI Research Engineer: Model Compression & Quantization

Tether; Barcelona; Híbrido; EUR 60.000 - 100.000; Jornada completa

Descripción de la vacante

Tether is seeking an AI Researcher to join our team in Barcelona to advance compression for multimodal AI systems. You will focus on reducing model footprints across LLMs and VLMs while maintaining high fidelity, enabling efficient edge deployment.

Responsibilities include designing and evaluating quantization, distillation, and pruning strategies, building robust pipelines, and reporting results. A strong background in transformers and PyTorch is required, with C++ familiarity a plus for

Formación

• Degree in Computer Science or related field.
• Experience with PyTorch deep learning frameworks.
• Hands-on experience with model quantization (QAT and PTQ).
• Research and hands-on experience with knowledge distillation.
• Experience with model pruning for compressing large models.
• Solid understanding of neural network architectures, especially transformers.
• Familiarity with C++ is a plus for low-level quantization kernels.

Responsabilidades

• Apply low-bit quantization to reduce model size and latency while preserving accuracy.
• Use knowledge distillation to transfer capabilities from larger teacher models to smaller ones.
• Implement pruning to remove redundant parameters and attention heads.
• Analyze trade-offs between model efficiency and accuracy; propose improvements.
• Research mixed-precision quantization and adaptive pruning schedules to optimize performance.
• Stay current with advances in model compression for multimodal AI architectures.
• Document methodologies, experiments, and results for reproducibility and collaboration.
• Publish findings in top-tier conferences to advance model compression field.

Conocimientos

PyTorch

Knowledge distillation

Model pruning

Transformers

C++ familiarity

Educación

Bachelor's degree in Computer Science

PhD preferred in NLP/ML

Herramientas

PyTorch

C++

Descripción del empleo

Tether is seeking an AI Researcher to join our team in Barcelona to advance compression for multimodal AI systems. You will focus on reducing model footprints across LLMs and VLMs while maintaining high fidelity, enabling efficient edge deployment.

Responsibilities include designing and evaluating quantization, distillation, and pruning strategies, building robust pipelines, and reporting results. A strong background in transformers and PyTorch is required, with C++ familiarity a plus for