Research Scientist
Google DeepMind
Posted 16 Nov 2026, 06:36 ET · Verified open on LinkedIn 16 Nov, 07:12 ET
This job at a glance
- Who can apply
- New grads and up to 3 years
- Must know
- StatisticsData structures & algorithms
- Nice to have
- LLMsMachine learning
- Work
- On-site in Cambridge
- Pay
- Not stated by the employer.
- Skill demand
- Machine learning28%Statistics23%LLMs10%Data structures & algorithms4%of 99 similar entry-level jobs posted in the US this week
- Listed
- Nov 16, 2026 · still open
Sign in free to see which of these skills your resume already shows.
About the Role
The People and AI Research (PAIR) team at Google DeepMind conducts foundational research to make human-AI partnerships productive, enjoyable, and fair. This multidisciplinary team of computer scientists, research engineers, social scientists, and designers invents novel interaction paradigms and develops open-source tools that shape how people interact with intelligent systems.
As a Research Scientist in Human-AI Interaction and Simulation, you will develop computational, cognitive, and statistical models of human behavior to understand, predict, and optimize how people interact with advanced AI systems. You will design empirical user studies, build predictive simulations of user interaction, and translate behavioral insights into architectural and UX improvements across Google DeepMind's frontier models.
Key Responsibilities
- Develop, validate, and land models of human behavior in the context of human-AI interactions, translating complex behavioral insights into actionable improvements for frontier AI models, interactive tools, and user interfaces.
- Focus on model integration and applied product impact across Google DeepMind's systems, pivoting focus from academic publishing to direct, applied contributions to frontier models.
- Conduct empirical experiments, build interactive research prototypes, and run simulations to evaluate human-agent collaboration, validate hypotheses, and refine research methodologies.
- Define data structures, simulation frameworks, experimental designs, and evaluation metrics for research solution development, establishing clear execution timelines and resource needs.
- Identify emerging research areas through internal and external collaborations, helping develop long-term strategies to expand Google DeepMind’s applied scientific impact.
Requirements & Tech Stack
- PhD in Computer Science, Artificial Intelligence, Cognitive Science, Computational Social Science, Human-Computer Interaction (HCI), or a related quantitative STEM discipline, or equivalent practical research experience.
- One or more first-author peer-reviewed scientific publications at academic venues (e.g., CHI, UIST, CSCW, NeurIPS, ICML) or public preprint repositories.
- Experience modeling and predicting human behavior when using interactive systems.
- Experience with core research methodologies, Human-Computer Interaction (HCI), and User Interface (UI) design.
- Experience with one or more of the following: computational cognitive modeling, simulated user co-pilots, human-agent collaboration, or Generative AI.
Preferred Qualifications
- 1 year of experience initiating, scoping, and executing independent research agendas or leading collaborative scientific projects in human-AI systems.
- Experience contributing to highly cited seminal papers, widely adopted open-source research frameworks, or recognized international scientific awards and fellowships.
- Experience in PAIR's core strategic focus areas: interactive code or music synthesis, LLM steerability, developer tools, Reinforcement Learning from Human Feedback (RLHF), and simulated human behavior alignment.
- Experience with interdisciplinary research bridging cognitive science, computer science, machine learning, and human-computer interaction (HCI).
- Experience building novel user interfaces, visualization tools, or sandbox environments to inspect, simulate, and evaluate human-agent interactions.