Skip to main content
AmigoFish

Google Cloud Tech

Google Cloud TechVideo

Format:
YouTube channel
New videos:
Daily
Typical length:
about 15 minutes, not counting Shorts
Latest video:
Not rated yet

About this show

Welcome to the official YouTube channel for Google Cloud’s technical community. Explore our tutorials, see exciting demos, stay up to date with product news, and find new developer tools helping you build what’s next.

Videos come from the channel's public YouTube feed. The artwork is the newest video thumbnail. Last updated . Names, artwork and text belong to their publishers.

Cosmos

Similar by topic and creator

See the full Cosmos

Videos

  • The Future of AI Infrastructure

    · 39 min

    Experience the future of AI infrastructure live on Nov 9 at 9 AM PT. Join Google SVP and Chief Technologist of AI and Infrastructure Amin Vahdat and Google Cloud VP and GM of AI and Computing Infrastructure Mark Lohmeyer on November 9 at 9:00 AM PT for a first look at AI infrastructure built for 24/7 AI agents. See how Google Cloud's AI Hypercomputer combines custom silicon, networking, and open software like PyTorch, vLLM, and Kubernetes to maximize performance. Resource links: * Event updates on Cloud OnAir → https://g.dev/ai/nov9 * Google Cloud AI Hypercomputer → https://g.dev/ai/ai-hypercomputer * Google Cloud TPUs → https://g.dev/ai/tpu #AIInfrastructure #GoogleCloud #AIHypercomputer #CloudTPU #AgenticAI Speakers: Amin Vahdat, Mark Lohmeyer, Maria Biggs Products Mentioned: Cloud TPUs, NVIDIA GPUs, Agent Compute Engine, GKE Inference Gateway

    Not rated yet
  • Fix your vibe coded app with one prompt

    · 17 min

    When you are vibe coding and fixing UI bugs one by one, it is easy to get stuck in an endless loop where fixing one thing breaks two others. In this episode of Build It, we tackle a frustrating set of UI bugs in our smart personal trainer app's active workout tracker, not by chasing isolated fixes, but by stepping back and defining a complete system. Instead of firing off piecemeal bug prompts, you will see how to map out presentational changes, component boundaries, page-level vs. exercise-level state variables (like current set and rest mode), display logic, and user actions to catch every edge case, including supersets and final set transitions. Then we use Gemini to generate a full end-to-end test suite and pass both the system spec and test cases to Google Antigravity in a single prompt so every scenario passes before we move on. Chapters: 0:00 - Intro: Why Systematic Updates Beat One-by-One Bug Fixes 1:06 - Identifying Friction and UI Bugs in the Active Workout Tracker 2:32 - Why Vibe Debugging Fails: Conflicting Fixes and Bug Whack-a-Mole 3:31 - Mapping Presentational Changes and Page Components 4:51 - Defining Page-Level vs. Exercise-Level State Variables 6:31 - Designing Display Logic and Catching Edge Cases 8:24 - Component-Level Display Logic: Current Set Card and Buttons 9:33 - Mapping User Action Logic and Rest Timer Transitions 10:45 - Generating an End-to-End Test Suite With Gemini 12:06 - Managing Antigravity Sessions 13:40 - Running the System Spec and Test Suite in One Prompt 14:33 - Live Verification: Rest Mode, Supersets, and Final Set Edge Cases More resources: Get the Prompts & Artifacts → https://g.dev/ai/build-it-resources Builder Central → https://g.dev/ai/builder-central Google Antigravity → https://g.dev/ai/build-it-with-agy Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #GoogleCloud #Gemini #Antigravity #NoCode #AIAgents #BuildIt Speakers: Devin Chasanoff, AI Builder Advocate Products Mentioned: Gemini, Google Antigravity, Google Cloud

    Not rated yet
  • Gemini Flash vs. Pro vs. Flash-Lite: How do you pick the right model for speed, cost, and reasoning?

    · 1 min

    A quick guide to choosing between Gemini Flash-Lite, Gemini Flash, and Gemini Pro based on latency, cost, and task complexity. Learn more: https://g.dev/ai/gemini-api-models (Note: We generated this explainer as part of an AI video experiment. Let us know in the comments if this quick format works for you.)

    Not rated yet
  • Grounding agents in your trusted semantic models using Knowledge Catalog

    · 4 min

    To learn more about Knowledge Catalog connectors → https://g.dev/cloud/connectors See how Google Cloud Knowledge Catalog automatically indexes semantic layers across LookML, BigQuery Graph, dbt MetricFlow, Databricks Unity Catalog, Snowflake, and SAP - giving AI agents the business context and semantics they need to deliver governed, accurate answers whilst saving cost and reduce latency. What along and see in this demo: * *Dynamic Context Graphs:* Visually explore how data assets, semantic models, schemas, joins, lineage, and governance connect across your organization. * *Side-by-Side AI Agent Comparison:* Watch a Knowledge Catalog grounded agent go head-to-head against a baseline control agent - producing accurate SQL with certified business logic in just 3 tool calls, cutting latency and token usage by ~67%. #Gemini #GoogleCloud #KnowledgeCatalog #AgentContext Speaker: Haengeun Chi Products Mentioned: Knowledge Catalog

    Not rated yet
  • Unlocking agentic workflows: Build agents with words

    · 31 sec

    Tired of configuring runtimes, model preferences, and rigid step-by-step workflow builders just to get an AI agent running? In this Google Cloud Live clip from Gemini at Work 2026, Adam Fisher (Director of Product Management, Google Cloud) explains how Gemini constructs the exact agents you need directly from natural language—and equips them with modular Skills to get the work done. In the full Google Cloud Live episode, we cover: * *One persistent cloud agent:* 24/7 continuity across desktop, mobile, Workspace, and CLI * *Zero copy enterprise data:* Grounding agents in live business context via Knowledge Catalog & Borderless Lakehouse * *Built in governance & smart routing:* Cryptographic identities, isolated sandboxes, and real-time spend caps #Shorts #GoogleCloud #GeminiAtWork #AIAgents #AgenticWorkflows Speaker: Adam Fisher Products Mentioned: Gemini

    Not rated yet
  • AI that knows your business: From data to ML fast

    · 31 sec

    Do you still need a specialized data engineering pipeline just to test a machine learning idea against your company's data? In this Google Cloud Live clip from Gemini at Work 2026, Gabe Weiss (Developer Advocacy Lead, Google Cloud) explains how grounding Gemini in your business context lets anyone sift through enterprise data, train an initial model, and prove a concept—so data engineers can take a validated prototype and run with it. In the full Google Cloud Live episode, we cover: * *Zero copy enterprise data:* Grounding agents in live schemas and business definitions via Knowledge Catalog & Borderless Lakehouse * *One persistent cloud agent:* 24/7 continuity across desktop, mobile, Workspace, and CLI * *Built in governance & smart routing:* Cryptographic identities, isolated sandboxes, and real-time spend caps #Shorts #GoogleCloud #GeminiAtWork #DataScience #AIAgents #KnowledgeCatalog Speaker: Gabe Weiss Products Mentioned: Gemini

    Not rated yet
  • 3 ways to design your agentic system

    · 9 min

    Build your agentic system with ADK → https://g.dev/cloud/cpm-vid-adk Learn three architectures for building modern AI applications: single agents, agent teams, and graph workflows. Using a real-world customer support example, Annie Wang breaks down when to rely on a flexible single agent, when to scale up to a specialized agent team, and how graph workflows support deterministic rules and provide reliability. Chapters: 00:00 - Introduction: 3 Ways to Build AI Apps 00:20 - The Customer Support Example 00:45 - Architecture 1: Single Agent 02:30 - Architecture 2: Agent Teams 04:00 - Architecture 3: Graph Workflows 05:30 - Connecting Agents, Teams, and Ordinary Code 06:45 - How to Choose the Right Architecture Speaker: Annie Wang Products Mentioned: Agent Development Kit

    Not rated yet
  • Why serverless GPUs are game-changing for AI development

    · 2 min

    Run Gemma on Cloud Run → https://goo.gle/4hAe9ko Follow AI inference best practices → https://goo.gle/4rC91kl Learn more about Cloud Run → https://goo.gle/4iU1qex Learn how Cloud Run Hackathon Finalist Abish Pius built “Dr. Cloud” — his take on an alternative to doomscrolling on WebMD. Chapters: 0:00 - Introduction 4:34 - The benefits of serverless GPUs 5:09 - Building Dr. Cloud 6:19 - Advice on how to get started 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #CloudRun #GoogleCloud #ServerlessGPUs Speaker: Abish Pius Products Mentioned: Cloud Run

    Not rated yet
  • Building a platform that can handle thousands of videos

    · 1 min

    Get started with Cloud Run → https://g.dev/cloud/starting-cloud-run Learn how Google Developer Expert Abhishek Sharma used Cloud Run to build one of India’s largest retailers, a platform that translates their training videos. Chapters: 0:00 - Intro 0:43 - Building with Cloud Run 1:00 - Autoscaling for thousands of videos 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #CloudRun #GoogleCloud #Gemini Speaker: Abhishek Sharma Products Mentioned: Cloud Run

    Not rated yet
  • How I built Battle Chess in one weekend… with no game dev experience

    · 3 min

    Play Battle Chess → https://g.dev/cloud/battle-chess Create your own Battle Chess theme → https://g.dev/cloud/create-your-own-battlechess Get started with Cloud Run → https://g.dev/cloud/get-started Nostalgic for classic 80s DOS games? Learn how Google Developer Expert John Capobianco vibe coded Battle Chess and how you can add your very own playable, custom theme. Chapters: 0:00 - Intro 0:40 - Recreating Battle Chess 1:15 - Vibe coding with Gemini 1:54 - Deploying with Cloud Run 2:34 - Advice for getting started 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #CloudRun #GoogleCloud #VibeCoding #Gemini Speaker: John Capobianco Products Mentioned: Cloud Run, Gemini

    Not rated yet
  • How I improved the UK Global Talent visa process

    · 1 min

    Check out Get Endorsed → https://g.dev/community/qualified-endorsed Get started with Cloud Run → https://g.dev/cloud/starting-cloud-run Google Developer Expert Omotayo Aina was tired of seeing talented applicants get rejected for UK Global Talent Visas due to preventable presentation mistakes. Using Cloud Run, he built GetEndorsed, so you can get your application reviewed without needing expensive immigration lawyers. Chapters: 0:00 - Intro 0:30 - Why I built it 1:10 - How it works 1:28 - Cost optimization with Cloud Run 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #CloudRun #GoogleCloud Speaker: Omotayo Aina Products Mentioned: Cloud Run

    Not rated yet
  • How I built my own English tutor

    · 2 min

    Try Zubro → https://g.dev/community/meet-zubru Learn more about building Zubro → https://g.dev/community/building-zubro Try Cloud Run → https://g.dev/cloud/starting-cloud-run Learn how Google Developer Expert Tomek Porozynski built Zubro, an AI-powered partner to help you practice English and give feedback on your grammar. Chapters: 0:00 - Intro 0:41 - Meet Zubro 1:22 - Building with Cloud Run 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #CloudRun #GoogleCloud #Gemini Speaker: Tomek Porozynski Products Mentioned: Cloud Run, Gemini, Secret Manager, Firebase

    Not rated yet
  • Real time vision for AI agents: Snapshots, streams & Gemini Live API

    · 7 min

    Code for this demo → https://goo.gle/vision-ai Learn how to build a multimodal AI agent with real-time vision using the Gemini Live API and Google Agent Development Kit (ADK). In this episode, we break down how camera frames enter a live voice session as image Blobs, when to send single snapshots vs. continuous video streams, and how to combine voice and vision without hurting latency or wasting tokens. What you'll learn: • End-to-end vision pipeline: Camera → Canvas → JPEG → Image Blob → Model • Snapshot vs. Stream: Why snapshots should be your default (and why streaming a browser breaks live voice) • Multimodal architecture: Combining live voice + camera input, plus the separate vision "watcher" pattern Chapters: 00:00 - Live Vision Demo & Episode Intro 00:37 - How an AI Agent Sees: Image Blobs in a Live Session 01:10 - End-to-End Frame Pipeline (Camera → Canvas → JPEG → Blob) 01:36 - 2 Myths: Live Video Files & "ADK Is Audio-Only" 02:21 - Snapshot vs. Stream: 2 Ways to Send Frames 03:25 - Token Costs & The Browser Streaming Trap 04:18 - Combining Live Voice & Real-Time Vision 05:26 - Advanced Pattern: Separate Vision "Watcher" Model 06:09 - Recap: Practical Rules for Multimodal Agent Vision 06:54 - Full Multimodal Agent Stack & Series Recap Preview What multimodal AI agent would you build? Share your ideas in the comments! More resources: Gemini Live API docs → https://g.dev/cloud/gemini-live Agent Development Kit (ADK) docs → https://g.dev/cloud/adk-docs Follow Annie online: LinkedIn → https://goo.gle/annie-linkedin X → https://goo.gle/annie-x Watch more The Omni App → https://g.dev/cloud/modern-ai-app 🔔 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #AIAgents #MultimodalAI #GeminiLiveAPI #ComputerVision #GoogleCloud #AgenticAI Speaker: Annie Wang Products Mentioned: Gemini Live API, Agent Development Kit (ADK), Google Gen AI SDK

    Not rated yet
  • Trusted agents, real outcomes: Driving AI impact at scale with Google's Agentic Data Cloud

    · 1 hr 6 min

    Moving from insights to autonomous action requires a data foundation built for the agentic era. Join Andi Gutmans and Google Cloud product leaders on November 4 at 9 AM PT for an inside look at how Google’s Agentic Data Cloud connects enterprise data to real-world outcomes. Discover the latest innovations across Google Cloud databases, analytics, and BI to power trusted agents, plus technical demos and customer stories. *Watch along and learn how:* * To elevate practitioners to shift from writing manual code to orchestrating high level business intent. * To ground agents in governed, real time business context so they can act with accuracy. * A borderless Lakehouse unifies data across multicloud, SaaS, and on-premises environments. 🔗 Resources: * Graph technologies → https://g.dev/ai/trusted-agentic-platform * Data strategy workshop → https://g.dev/ai/make-you-data-ai-ready * What is borderless Lakehouse? → https://g.dev/cloud/lakehouse-introduction * Google Data Cloud Graph eBook → https://g.dev/cloud/unlock-graph-analytics * Practical guide to AI with BigQuery → https://g.dev/ai/bigquery-ai-ebook * Request a Looker demo → https://g.dev/cloud/better-data-analysis * Looker trial offer → https://g.dev/cloud/try-looker * Spanner Omni → https://g.dev/cloud/deploy-spanner-anywhere * Spanner queues → https://g.dev/cloud/transactional-messaging-agentic-workloads * Oracle AI Database @ Google Cloud → https://g.dev/ai/agentic-enterprise * Cloud SQL for MySQL → https://g.dev/cloud/blue-green-deployments * PostgreSQL for agents → https://g.dev/ai/real-time-data-agent-scale Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #DataAnalytics, #Databases, #BusinessIntelligence, #AgenticDataCloud, #GoogleCloud Speakers: Andi Gutmans, Sean Zinsmeister, Gabe Weiss, Mikul Bhatt, Ryan Polivka Products Mentioned: Gemini, BigQuery, Knowledge Catalog, Spanner, Borderless Lakehouse, Spanner Graph, Looker, Conversational Analytics

    Not rated yet
  • Cloud Load Balancing explained: How does global traffic routing work behind a single IP address?

    · 1 min

    How global anycast routing and Cloud Load Balancing distribute user traffic to the closest healthy backend without a single point of failure. Learn more: https://g.dev/cloud/load-balancing-overview (Note: We generated this explainer as part of an AI video experiment. Let us know in the comments if this quick format works for you.)

    Not rated yet
  • I built an agent to run my YouTube channel

    · 1 hr 38 min

    Most AI agents fail in production because a single prompt can’t guarantee order, safety checks, or human approval. Today, we’re building a fully autonomous, multimodal AI video studio from scratch using Google’s Agent Development Kit. Imagine running a daily cozy-spooky video channel ("VibeTube"). Every night, your audience sends you chaotic ideas: voice notes, porch videos, spooky photos, and text comments. Gemini can read them, Imagen can draw the posters, and Veo can render the video. The problem? Generative models can do the chores, but they can’t run the studio. They can't refuse a bad idea before it burns your API budget, pause for your approval, or wait for a slow video render without timing out. A prompt can’t promise order. A workflow graph can. In this live build, we are walking through the VibeStudio (Haunted Edition), architecture to turn a chaotic creative process into a deterministic, production ready ADK Workflow on Google Cloud. *What the team is building live:* * 🧠 Beyond basic prompts: Moving from single turn agents to structured ADK Workflow graphs (fan-out, parallel JoinNodes, and smart routers). * 👁️ Multimodal RAG ingestion: Using Gemini to transcribe voice notes, caption photos, and watch video clips—feeding them all into a single Vertex AI RAG Engine corpus. * 🛑 The 2 types of suspension: How to make your agent pause for a human-in-the-loop decision and asynchronously wait on long running Veo video renders without holding connections open. * 💾 Agent memory: Wiring up Memory Bank callbacks so your studio actually remembers what it did yesterday. * 🚀 Prompt to production: Deploying the entire VibeStudio workbench to Cloud Run and publishing our first AI generated short. Resources: 👉 Follow along with the Codelab → https://g.dev/ai/vibe-studio 💻 GitHub repo → https://g.dev/ai/vibe-studio-multimodal Speakers: Annie Wang, Logan Hennessy Products Mentioned: Agent Development Kit, Veo

    Not rated yet
  • Stop vibe coding blind: Spec driven UI design with AI agents

    · 11 min

    Great apps need more than a working back end, they need a thoughtful user experience. Instead of jumping straight into vibe coding and chasing individual features, this episode of Build It walks through creating a complete front-end design spec for our smart personal trainer app with zero coding required. Rather than asking an AI model to generate the entire spec in one shot, you will see how to prompt Gemini in Google Antigravity to build a custom two-agent team: a conversational Planning Companion that acts as a thinking partner and explains front-end concepts in plain English, paired with a dedicated Scribe agent that documents every finalized decision. Follow along as we turn prototype screenshots from Google AI Studio into a clear visual roadmap covering pages, component layouts, user flows, and natural language test cases. Chapters: 0:00 - Intro: Why front end design specs beat vibe coding 1:19 - Prompting Gemini to build a custom agent team 2:57 - Designing an AI thinking partner and teaching companion 4:47 - Separating the planning companion from the scribe agent 5:43 - Starting a fresh session with the client planning agent 6:49 - Turning AI Studio prototype screenshots into a UI roadmap 7:37 - Learning front end concepts play-by-play (pages and components) 9:00 - Applying your design taste without the mental overhead 9:40 - Reviewing the complete front end design spec 🔗 Resources: * Get the prompts & artifacts → https://g.dev/ai/build-it-resources * Builder central → https://g.dev/ai/builder-central * Google Antigravity → https://g.dev/ai/build-it-with-agy Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech #GoogleCloud #Gemini #Antigravity #NoCode #AIAgents #BuildIt Speakers: Devin Chasanoff, AI Builder Advocate Products Mentioned: Gemini, Google Antigravity, Google AI Studio, Google Cloud

    Not rated yet
  • More AI agents = better results?

    · 58 sec

    Do more AI agents always mean better results? Not always, it depends on what you're building and how your agents work together. When building an app with AI, adding more agents comes with real trade-offs: • Majority voting (asking multiple agents to fix the same bug) → helps surface consensus, but the most popular answer can still be wrong • Agent teams (splitting UI, data, and testing across specialized agents) → can run tasks in parallel to finish faster and catch missed bugs • Sequential handoffs (agents building on each other's work) → one early mistake can cascade down the chain An agent team can help you explore ideas, move faster, and catch bugs, just remember that more agents doesn't automatically mean a better answer. #AIAgents #MultiAgentSystems #AIDevelopment #SoftwareEngineering #LLM #GoogleCloud Speaker: Annie Wang

    Not rated yet
  • Cloud Run vs.Compute Engine: When should you run a container on serverless vs. managing a full VM?

    · 1 min

    How to decide between deploying containers on serverless Cloud Run versus managing your own virtual machines on Compute Engine. Learn more: https://g.dev/cloud/what-is-cloud-run (Note: We generated this explainer as part of an AI video experiment. Let us know in the comments if this quick format works for you.)

    Not rated yet
  • BigQuery vs. Production SQL: When should you use a data warehouse instead of a relational database?

    · 1 min

    Why analytical queries slow down a production transactional database, and when to use a columnar data warehouse like BigQuery instead. Learn more: https://cloud.google.com/bigquery/docs/introduction (Note: We generated this explainer as part of an AI video experiment. Let us know in the comments if this quick format works for you.)

    Not rated yet

More in Technology

See all Technology shows