Research Scientist, Multimodal Alignment, Safety and Fairness, DeepMind

What you'll do

• Generate research ideas and design, implement, and evaluate methods, interfaces, and tools for sociotechnical AI systems.
• Present research findings internally and externally, and collaborate across teams while contributing individually.
• Own substantial technical projects from ideation and design through implementation and evaluation.

What they ask for

• PhD in Computer Science, Machine Learning, or a related technical field, or equivalent practical experience.
• Experience with Python and deep learning frameworks (e.g., JAX, Flax, or Gemma).
• Experience developing multimodal AI models and systems, and conducting research and development including experimental design, implementation, and analysis.
• Publication record in machine learning conferences (e.g., NeurIPS, CVPR, ICML, ICLR, ICCV, ECCV).

Nice to have

• Experience fine-tuning and post-training LLMs using Reinforcement Learning (RL) or other alignment methods.
• Experience developing generative AI architectures, techniques, and agentic AI solutions; working with large-scale vision-language models (VLMs) is also preferred.

Pay

$147000 - $210000 USD base salary

Tools

Python, JAX, Flax, Gemma, Reinforcement Learning (RL)

AI focus

Research and development of multimodal AI systems, including alignment, safety, and fairness; preferred work includes LLM post-training, generative and agentic AI, and vision-language models.