Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore how generative AI is transforming the pathway from research breakthroughs to real-world products in this 24-minute conference talk by Xavi Amatriain, VP of Product, AI and compute enablement at Google. Discover the fundamental shift from the "Age of Data" to an era of rapidly evolving models, emergent agentic capabilities, and "superhuman data" that presents both unprecedented opportunities and unique challenges for product and engineering teams. Learn about the compressed research-to-product lifecycle enabled by powerful foundation models and the new strategies required for rapid iteration and forward-thinking development. Gain actionable insights for translating cutting-edge research—including novel architectures, advanced prompting techniques, and agentic systems—into scalable, reliable product features. Understand evolving best practices for robust evaluation beyond traditional metrics, techniques for mitigating hallucinations, approaches to responsible AI development, and the changing role of data in modern AI systems. Examine the increasing importance of comprehensive tools and platforms for managing experimentation, ensuring reproducibility, and accelerating the journey from concept to production-ready AI. See how Google's full-stack infrastructure, from custom silicon to advanced frameworks, exemplifies delivering models at the Pareto frontier while ensuring production reliability, scale, and cost efficiency. The presentation covers the evolution from linear models to deep learning, the acceleration brought by generative AI, multimodal agents like Astra, open source tooling, agent hackathons, and future applications in science and the physical world.
