This resource provides insights and tools from a research scientist specializing in artificial intelligence and 3D computer vision, aimed at professionals and academics in AI, healthcare, and technology fields. It offers practical knowledge and methodologies for building secure and effective AI systems.

• Watermarking and Model Security: Developed SynthID, a solution for watermarking AI models to protect intellectual property and ensure traceability in deployments.
• Medical AI Agents: Created Med-PaLM and Med-Gemini, advanced language models and diagnostic dialogue systems tailored for healthcare applications, improving medical decision-making.
• Uncertainty Estimation: Implements conformal and Bayesian methods to assess uncertainty in deep learning and large language models, enhancing reliability in critical tasks like natural language inference.
• 3D Computer Vision: Works on generative models for 3D shapes and point clouds, along with classical techniques for segmentation and object detection, supporting applications in visualization and analysis.
• Cross-Organizational Projects: Leads research initiatives that result in deployments and publications, including submissions to Nature, demonstrating impact in both academic and industrial settings.