Research Scientist, Toronto

Snapshot

The Google DeepMind Toronto team is dedicated to driving fundamental research breakthroughs across key areas of Artificial Intelligence. Our work spans multiple modalities (text, image, video), with core research efforts focused on advancing Large Language Models (LLMs), text-to-image and text-to-video generation, intelligent agents, and foundational representation learning. As a Research Scientist, you will contribute to discovery by exploring novel algorithms, developing new theories, and pushing the frontiers of AI. We foster a highly collaborative, supportive, and dedicated environment where ambitious scientific exploration thrives.

About Us

Artificial Intelligence could be one of humanity’s most useful inventions. At Google DeepMind, we’re a team of scientists, engineers, machine learning experts and more, working together to advance the state of the art in artificial intelligence. We use our technologies for widespread public benefit and scientific discovery, and collaborate with others on critical challenges, ensuring safety and ethics are the highest priority.

The Role

Research Scientists at Google DeepMind lead our efforts in developing novel algorithmic and data solutions towards the end goal of solving and building Artificial General Intelligence.

As a Research Scientist on our team, you will tackle fundamental challenges in enabling AI systems to learn, remember, and adapt over time, mirroring capabilities essential for true intelligence. Your work will center on advancing memory and adaptation techniques specifically for multimodal large language models (MLLMs). We are exploring the frontiers of long-term and short-term memory systems, investigating how procedural memory can be effectively integrated, and developing novel approaches to continual learning (including supervised, reinforcement, and in-context methods). A core focus will be mitigating catastrophic forgetting to ensure enduring knowledge and skill retention. Through this research, we aim to create agents capable of persistent operation and reliable task execution in complex, dynamic environments over long timescales, a crucial step towards achieving AGI.

Key responsibilities:

• Lead the conceptualization, design, and implementation of novel algorithms, models, and architectures focused on memory systems (long-term, short-term, procedural) within multimodal large language models (MLLMs) or other advanced AI systems.
• Develop, investigate, and push the boundaries of continual learning techniques (including supervised learning, reinforcement learning, in-context learning, and potentially new paradigms) to enable robust adaptation over long timescales.
• Formulate and experimentally validate innovative strategies to explicitly address and mitigate catastrophic forgetting in learning systems.
• Design, implement, and refine rigorous evaluation methodologies, metrics, and benchmarks specifically tailored to assess memory efficacy, adaptation speed, knowledge retention, and long-term task performance in complex, dynamic environments.
• Plan, execute, and analyze large-scale experiments to test fundamental hypotheses, identify limitations of current approaches, and demonstrate breakthroughs in building persistent and adaptive AI agents.
• Drive foundational research directions and contribute significantly to the scientific understanding of memory, adaptation, and lifelong learning as components of Artificial General Intelligence.
• Collaborate closely with other researchers and engineers across Google DeepMind, sharing insights, leveraging diverse expertise, and contributing to the collective mission.
• Publish impactful research findings in leading peer-reviewed AI conferences and journals, advancing the state-of-the-art in the field.

About You

In order to set you up for success as a Research Scientist at Google DeepMind,  we look for the following skills and experience:

• PhD in Computer Science, a related technical field, or equivalent practical experience demonstrating research expertise.
• Experience in core Machine Learning algorithms, deep learning, and research, particularly in areas relevant to large language models (LLMs), sequence modeling, or multimodal learning.
• Proficiency in Python and experience with major deep learning frameworks (e.g., PyTorch, JAX, TensorFlow).
• A track record of publications in top-tier Machine Learning / AI conferences or journals (e.g., NeurIPS, ICML, ICLR, EMNLP, CVPR, AAAI, etc.), demonstrating impactful contributions.
• Strong software engineering abilities with proven experience in designing and executing complex experiments ("hands-on" research).

In addition, the following would be an advantage: 

• Demonstrated research experience and publications specifically in one or more of the following areas:
• Memory, long-context processing, and information retrieval (e.g., RAG) in large models.
• Model adaptation and lifelong learning (e.g., fine-tuning, ICL, continual learning, RL-based adaptation).
• Multimodal representation learning and generation (text, image, video).
• Experience designing and implementing novel evaluation methodologies for AI capabilities.
• Experience with large-scale model training and distributed computing environments.
• Experience working effectively in a collaborative research environment (industry or academic lab).
• Experience initiating, leading, or significantly contributing to ambitious research projects.

Applications close: Thursday 1st May. 

At Google DeepMind, we value diversity of experience, knowledge, backgrounds and perspectives and harness these qualities to create extraordinary impact. We are committed to equal employment opportunity regardless of sex, race, religion or belief, ethnic or national origin, disability, age, citizenship, marital, domestic or civil partnership status, sexual orientation, gender identity, pregnancy, or related condition (including breastfeeding) or any other basis as protected by applicable law. If you have a disability or additional need that requires accommodation, please do not hesitate to let us know.