Google I/O extended 2024 - What's new in Cloud?

Transcript

2. • The GDG Cloud Montreal team • Google Cloud Next ‘24 TL;DR • Some Google Workspace Love • What’s new in Google Cloud • Next GDG Cloud Montreal event Google I/O Extended - Speaker slides
3. The GDG Cloud Montreal Team
4. GDG Cloud Montréal Meet the team! Stacy Véronneau Director of Cloud Solutions and Sales Engineering Florian Ley Application Manager
5. Next ‘24 TL;DR
6. Google Cloud Next '24, held in Las Vegas from April 9-11, 2024, was a major tech conference showcasing Google Cloud's latest innovations and advancements in cloud computing, artificial intelligence, data analytics, cybersecurity, and more. Hello Sphere! Next ‘24 - TL;DR
7. What happens in Vegas when it comes to Google Cloud, can’t stay in Vegas! Up to 30 000 people attended and the event will be back in Vegas in 2025 around the same time. What Happens In Vegas… Next ‘24 - TL;DR
8. Generative AI advancements: Google announced major updates to its Gemini AI models, including the public preview of Gemini 1.5 Pro, enhanced capabilities for coding and threat intelligence, and integrations with Google Workspace. Custom silicon and infrastructure enhancements: Google revealed new custom chips like the TPU v5e and Google Axion, along with improvements to its AI Hypercomputer architecture. Expanded data and AI ecosystem: Google emphasized its commitment to partnering with leading companies across the AI stack to offer customers a wide range of solutions and choices. Focus on developer productivity and collaboration: Google introduced new tools like Gemini Code Assist to help developers write code more efficiently, and highlighted collaborative features within Google Workspace. Key highlights from Next '24 included: Next ‘24 - TL;DR
9. Some Workspace Love
10. • Gemini 1.5 Pro: A more advanced version of Google's AI model for Workspace, offering enhanced capabilities in text, image, audio, and video understanding. • AI Workflows: Streamlining tasks within Workspace by using AI to automate repetitive actions and suggest relevant information. • Side Panel Access: A new way to interact with Gemini AI features directly within Workspace applications. • Gemini in Google Meet: AI-powered features like advanced meeting summaries and Q&A for improved engagement. • AI-Powered Virtual Assistants in Google Chat: Providing personalized support and task automation in conversations. Some Workspace Love What’s new in Workspace
11. What’s new in Google Cloud
12. Google's advanced Gemini models are now available on Google Cloud. Gemini 1.5 Pro offers powerful performance across tasks, while the Flash version prioritizes speed and efficiency for large-scale applications. Example: A news organization uses Gemini 1.5 Pro to generate summaries of long articles, while a social media platform employs Gemini 1.5 Flash to analyze millions of posts in real-time for sentiment analysis. Gemini 1.5 Pro and Flash What’s new in GCP
13. Accelerate the building of AI applications Example: Build a generative AI application using the Gemini API and Vertex AI, and deploy it rapidly using Cloud Run or GKE, so you can focus on what's core to your business and leave the auto-scaling to Google. • Cloud Run | GKE for deploying the applications • Cloud SQL for PostgreSQL as a vector database • Vertex AI for embeddings model and LLM Benefits: The benefits of using Vertex AI and those tools for generative AI applications • Makes it easier to build and deploy generative AI applications • Provides a secure and scalable solution • Offers observability to monitor the application's performance Vertex AI + Cloud Run | GKE | Cloud SQL What’s new in GCP
15. The service formerly known as (Generative AI App Builder then Vertex AI Search and Conversation) • No-code interface: Enables you to build complex conversational agents without writing a single line of code, making it accessible to a wider audience. • Generative AI capabilities: Utilizes state-of-the-art language models for natural and engaging conversations. • Integrations with other Google Cloud services: Seamlessly connects with Vertex AI Search, Vertex AI Conversation, and other tools in the Google Cloud ecosystem. • Customization and extensibility: Allows you to tailor the agents to specific use cases and industries. • Conversational AI Flows: Create dynamic dialogues with branching logic and conditions based on user inputs. • Knowledge connectors: Connect your agents to various knowledge bases, documents, or APIs to access and retrieve relevant information during conversations. • Multiple deployment options: Deploy your agents on various platforms, including websites, mobile apps, messaging platforms, and contact centers. Vertex AI Agents What’s new in GCP
16. Examples: • Customer support: Provide instant and accurate responses to customer queries, resolve issues, and automate common tasks. • Sales and marketing: Engage with potential customers, qualify leads, and provide personalized recommendations. • HR and onboarding: Automate employee onboarding, answer HR-related questions, and assist with training and development. • Education and training: Deliver interactive learning experiences, provide tutoring and feedback, and assess student progress. • Healthcare: Offer virtual consultations, schedule appointments, and provide information about medical conditions and treatments. Vertex AI Agents What’s new in GCP
17. Vertex AI Agents - App type What’s new in GCP
18. Vertex AI Agents - Data Sources What’s new in GCP
19. What are grounding systems? • A way to connect the model's output to real-world information • Uses retrieval augmented generation (RAG) • Takes user queries, retrieves relevant information from a data source, and then uses that information to inform the large language model's response Benefits of using Vertex AI Search: • Simpler than building your own Grounding system from scratch • Stores and searches the information you want to use to ground your large language model RAG and Vertex Search What’s new in GCP
20. • One-Stop Shop: Simplify your AI journey with a comprehensive platform that caters to every stage of development. • Innovation at Your Fingertips: Stay ahead of the curve with access to the latest and greatest AI models. • Built for the Future: Create applications that leverage the full potential of AI, today and tomorrow. • Google's AI Expertise: Benefit from Google's world-class research and infrastructure. Trust in the quality and scalability of the models you deploy. • Beyond Text: Explore multimodal models that combine text, images, and other data types for truly innovative applications. Vertex AI Model Garden What’s new in GCP
21. Next GDG Cloud Montreal Event
22. Next GDG Cloud Montreal Event In this talk we introduce Vertex AI Pipelines to train image classification models on tpus. We will be using NNX a Google DeepMind high level API that is at the backbone of many vision and multimodal models at Google and OpenSource, we will do many pipelines examples with the objective of building a TPU training pipeline using Kubeflow and Vertex. This session is a good intro for practitioners and students interested in numerical and scientific computing and MLOps. July 10th at Noon online, a talk by David Cardozo from Updata around MLOPs and TPU’s on GCP.
23. Thank You Stacy Véronneau Director of Cloud Solutions and Sales Engineering -SE Cloud Experts-