AI Research Engineer (Model Compression & Quantization)

Tether.io; Netherlands; On-site; EUR 90,000 - 130,000; Full time

Job summary

Tether.io is seeking an experienced AI researcher to advance compression for multimodal systems, including LLMs and VLMs. You will reduce model footprint and compute cost while preserving accuracy across text, image, and audio streams.

You will lead quantization, distillation, and pruning efforts, publish findings, and collaborate across teams to push state-of-the-art in model efficiency for edge deployment.

Qualifications

• PhD preferred in NLP or ML with strong publication record.
• Hands-on experience with PyTorch and ML research.
• Experience quantizing and compressing large models.
• Familiarity with distillation and pruning approaches.

Responsibilities

• Apply low-bit quantization to reduce model size and latency for multimodal AI.
• Leverage knowledge distillation to transfer capabilities to smaller models.
• Implement pruning to remove redundant parameters and attention heads.
• Analyze trade-offs between efficiency and accuracy across methods.
• Research mixed-precision quantization and adaptive pruning schedules.
• Stay current with model compression research for multimodal architectures.
• Document methodologies and results for reproducibility and collaboration.
• Author technical papers and publish findings in top conferences.

Skills

Publications A* conferences

AI R&D experience

NLP/ML knowledge

Education

PhD in NLP

PhD in ML

Tools

PyTorch

C++

Quantization tooling

Job description

About The Job

As a member of our AI research team, you will drive innovation in model compression and efficient deployment for advanced multimodal AI systems, including large language models (LLMs) and vision-language models (VLMs). Your work will focus on reducing model footprint and computational cost while preserving accuracy, enabling high-performance AI to run efficiently across resource‑constrained edge devices. You will apply and advance compression techniques such as quantization, knowledge distillation, and pruning to streamline complex multimodal architectures that integrate text, images, and audio.

Responsibilities

• Apply low‑bit quantization to reduce model size and inference latency for generative AI models (LLMs, VLMs, multimodal) while maintaining accuracy and output quality.
• Leverage knowledge distillation to transfer capabilities from larger teacher models to smaller student models, enabling efficient multimodal reasoning across text, image, and audio inputs.
• Implement pruning techniques to remove redundant parameters and attention heads, reducing computational overhead without sacrificing task performance.
• Analyze trade‑offs between model efficiency (size, latency, memory) and accuracy across quantization, distillation, and pruning methods; propose improvements based on empirical findings.
• Research and apply mixed‑precision quantization and other advanced compression strategies (e.g., adaptive pruning schedules, distillation with intermediate feature matching) to optimize the accuracy–performance balance.
• Stay current with the latest research in model compression, including emerging techniques for multimodal and generative architectures.
• Document methodologies, experiments, and results clearly to support reproducibility, internal collaboration, and stakeholder communication.
• Author technical papers and publish findings in top‑tier conferences (NeurIPS, ICML, ICLR, CVPR, ACL, AAAI) to advance the field of model compression for multimodal AI.

Qualifications

• Degree in Computer Science or related field; Ph.D. in NLP, Machine Learning, or a related field preferred.
• Solid track record in AI R&D with publications in A* conferences.
• Experience with PyTorch deep learning frameworks or equivalent.
• Hands‑on experience with model quantization (both QAT and PTQ).
• Hands‑on experience with knowledge distillation for compressing large models into smaller, efficient ones.
• Hands‑on experience with model pruning for compressing large models into smaller, efficient ones.
• Solid understanding of neural network architectures and training processes, including transformers (LLMs, VLMs), backpropagation, optimization, and fine‑tuning.
• Familiarity with C++ is a plus, especially for implementing low‑level quantization kernels or inference optimizations.