Research Scientist - Generative AI & Multi-Modal Systems

WeAreTechWomen; Greater London; On-site; GBP 120,000 - 180,000; Full time

Job summary

Google DeepMind is seeking a Research Scientist to design large-scale experiments, prototype new architectures, and publish findings across multi-modal AI problems. You will drive end-to-end research from ideas to deployment with independence and collaboration across international teams.

The role emphasizes literature review, rigorous experimentation, and delivering impactful results for real-world applications in a fast-paced environment.

Qualifications

• PhD in computer science, mathematics, applied statistics, ML or equivalent practical experience.
• Experience training generative models (LLMs, image, video).
• Experience with ML frameworks (TensorFlow, JAX, PyTorch).
• Experience conducting research.
• Publications in AI conferences (NeurIPS, CVPR, ICCV, ICLR).

Responsibilities

• Optimize model architectures and distributed training pipelines across accelerators (TPUs/GPUs).
• Design and train state-of-the-art generative models across image, video, and audio domains.
• Develop and deploy reinforcement learning algorithms to align model behaviors and improve reasoning.
• Drive efficient inference via quantization, pruning, distillation, decoding, and kernel optimization.
• Bridge research and production by translating breakthroughs into scalable algorithms.

Skills

PhD or equivalent experience

Generative models experience

TensorFlow/JAX/PyTorch

Research experience

Publications in AI conferences

Education

PhD degree in CS/ML or related

Tools

TensorFlow

JAX

PyTorch

Job description

Google DeepMind is seeking a Research Scientist to design large-scale experiments, prototype new architectures, and publish findings across multi-modal AI problems. You will drive end-to-end research from ideas to deployment with independence and collaboration across international teams.

The role emphasizes literature review, rigorous experimentation, and delivering impactful results for real-world applications in a fast-paced environment.