Research Scientist

Our special interdisciplinary team combines the best techniques from deep learning and reinforcement learning to build general-purpose learning algorithms and apply these to medicine and healthcare. We have already made a number of high profile breakthroughs towards building artificial general intelligence for medicine, and we have all the ingredients in place to make further significant progress over the coming years to real world impact. In this role, you will accelerate the development and clinical deployment of our frontier multimodal AI medical assistant and co-clinician systems. You will drive forward research toward a safe, medically licensed agentic system for triadic care, deployed to doctors and health systems globally (spanning key collaborations with world-leading researchers and clinicians in the US, EMEA and APAC). Artificial intelligence will be one of humanity’s most transformative inventions. At Google DeepMind, we are a pioneering AI lab with exceptional interdisciplinary teams focused on advancing AI development to solve complex global challenges and accelerate high-quality product innovation for billions of users. We use our technologies for widespread public benefit and scientific discovery, ensuring safety and ethics are always our highest priority. We are pushing the boundaries across multiple domains. Our global teams offer diverse learning opportunities and varied career pathways for those driven to achieve exceptional results through collective effort. Individual pay is determined by factors including job-related skills, experience, and relevant education or training. US: $207000 - $300000 (USD) + 20% bonus target + equity + benefits Learn more about benefits at Google . Implement experiments across both model training and system harness to advance a state-of-art AI co-clinician. Drive research on clinical reasoning and longitudinal disease management, reliable agentic action-taking and automation, and verifiable risk management with continual adaptation. Define milestones and build simulation-to-real evaluation frameworks measuring clinical capability, safety, and efficiency. Validate hypotheses in-silico and in-vivo with precise clinical evaluations, partner with domain experts for real-world prospective validation. Collaborate closely with fellow researchers and help build and scale the long-term research team. Minimum qualifications: PhD in Computer Science, Artificial Intelligence, or a related quantitative field, or equivalent practical experience. 4 years of experience in LLM fine-tuning/Reinforcement Learning, autonomous and human-on-loop agent development, and designing evaluation frameworks for real-world applications. Experience providing technical leadership to execute research from concept to product. Experience with clinical and healthcare workflows, data, and environments. Experience partnering with vendors for data and rubric curation in real-world domains, and generating synthetic data and rubrics. Preferred qualifications: Experience building agentic benchmarks involving reasoning-driven API read/write, computer control, or healthcare record integrations Experience with training agents for long-horizon tasks, computer control, domain safety, uncertainty calibration, and human-on-the-loop systems. Experience in full-lifecycle LLM development, deploying autonomous/human-on-loop agents in safety-critical real-world environments. Track record of publishing in AI venues (e.g., NeurIPS, ICML, ICLR), or scientific/medical journals (e.g., Nature). PhD in Computer Science, Artificial Intelligence, or a related quantitative field, or equivalent practical experience. 4 years of experience in LLM fine-tuning/Reinforcement Learning, autonomous and human-on-loop agent development, and designing evaluation frameworks for real-world applications. Experience providing technical leadership to execute research from concept to product. Experience with clinical and healthcare workflows, data, and environments. Experience partnering with vendors for data and rubric curation in real-world domains, and generating synthetic data and rubrics.