Research Scientist, Frontier Ai, Deepmind

Google Google · Big Tech · Zürich, Switzerland

Research Scientist at Google DeepMind focused on advancing the frontiers of artificial intelligence, specifically developing next-generation multimodal models. This role involves defining and leading ambitious research projects, designing innovative model architectures, and contributing to the overall research strategy with a strong emphasis on publications in top ML conferences.

What you'd actually do

  1. Define and lead ambitious research projects, setting the technical direction and ensuring the successful execution of the research agenda.
  2. Design and implement innovative model architectures that effectively integrate diverse datasets, particularly across image and text modalities.
  3. Develop and establish benchmarks for evaluating multimodal models across a wide range of tasks and domains.
  4. Contribute to the overall research strategy of the team, identifying new research directions and influencing the broader research community through publications and presentations.

Skills

Required

  • PhD in Machine Learning, Computer Science, Statistics or a related field, or equivalent practical experience
  • 15 years of experience leading a research agenda
  • Experience in deep learning
  • Experience with Python
  • Experience with neural network training
  • Experience training large multimodal models
  • Experience evaluating large multimodal models
  • Experience interpreting large multimodal models
  • Experience training large language models
  • Experience evaluating large language models
  • Experience interpreting large language models
  • A publication record in machine learning conferences (e.g., NeurIPS, CVPR, ICML, ICLR, ICCV, ECCV)

Nice to have

  • Experience in research leadership across organizations
  • Experience leading a diverse set of research scientists, research engineers, and software engineers

What the JD emphasized

  • 15 years of experience leading a research agenda
  • A publication record in machine learning conferences (e.g., NeurIPS, CVPR, ICML, ICLR, ICCV, ECCV)

Other signals

  • frontier research
  • multimodal models
  • deep learning
  • large language models
  • publication record