🎉 And that’s a wrap on #DevFest 2025! A massive thank you to all the developers, organizers, and speakers who came together to learn, connect, and grow. This season was packed with achievements. Together, you were busy: 🛠️ Participating in countless interactive workshops 💻 Contributing to open-source projects 💡 Sharing solutions to complex technical challenges We appreciate you being a part of this global community. See you for DevFest 2026! 👋 Find a Google Developer Group near you → https://goo.gle/4a9S0ax
Google for Developers
Technology, Information and Internet
Mountain View, CA 3,499,130 followers
Join a community of creative developers and learn how to use the latest in technology—from AI and cloud, to mobile & web
About us
Discover the latest technologies, resources, events, and announcements to help you build smarter and ship faster. Explore more at developers.google.com
- Website
-
http://developers.google.com
External link for Google for Developers
- Industry
- Technology, Information and Internet
- Company size
- 10,001+ employees
- Headquarters
- Mountain View, CA
- Specialties
- coding, engineering, firebase, android, cloud, web development, and mobile development
Updates
-
Transform your daily workflows at no cost. ✨ We partnered with DeepLearning.ai to launch a new course: "Gemini CLI: Code & Create with an Open-Source Agent." Learn how to integrate Gemini CLI, leverage memory features for context-aware AI, and extend capabilities with MCP servers. Start learning → https://goo.gle/4aa5CSu
-
ADK TypeScript is now available for all JavaScript and TypeScript developers. Watch the tutorial below to see how a code-first approach simplifies agent orchestration ✨
ADK TypeScript is here. This toolkit brings a code-first approach to agent development. Use it to define agent logic, tools, and orchestration directly in TypeScript with full type safety. In this tutorial, Mandy Chan shows how to build your first AI agent using the new open-source toolkit with two demos to get you started: Science Teacher Agent (04:50) and Customer Order Agent (08:30) Check out these resources to start building ⬇️ Blog: https://goo.gle/45wY3CU Quickstart guide: https://goo.gle/4qOVV24 Github Samples: https://goo.gle/3Z4clHs Documentation: https://goo.gle/4bOlbAE
-
Great software is rarely built in a silo. 🧱 While individual coding speed is valuable, long-term velocity comes from shared context. By treating pair programming and code reviews as active knowledge transfer, rather than just administrative hurdles, teams can tackle complexities that are impossible to solve alone. How does your team collaborate? 🧠
-
-
Optimize your analytics workflows with NVIDIA GPUs on Google Cloud → https://goo.gle/4jZKUs8 A new learning pathway in partnership with NVIDIA is now available for data enthusiasts. ✨
-
-
Google for Developers reposted this
Convincing a customer to trust in a new technology is one thing; convincing over the top developers in the Google Developer Experts (GDE) program is another. However, the #Chrome Prompt API made that transition seamless. By leveraging Gemini Nano for local inference in the #Chrome Browser (#WebAI), we pushed processing to the edge. In Browser Prompt API handled safety-critical decisions, processing 10Hz GPS telemetry with sub-50ms latency to determine track position in real-time. When speed is non-negotiable, in Browser local AI delivers. CC: Jason Mayes, Google Developer Experts, Alvaro Huanca, Karen Acosta Parra
AI agents in a Jupyter notebook are safe. AI agents on a race track are a different story. For the High-Velocity AI Field Test, we issued a challenge to the Google Developer Experts (GDEs): Leverage the full Google stack to build a "Trustable AI" system that meets or exceeds the current state-of-the-art. The mission was to provide a driver with verifiable, real-time intelligence in a high-stakes environment where cloud latency isn't just a bug—it's a safety risk. The "Aha" Moment: The Human in the Loop - We quickly realized that trust doesn't come from the model alone; it comes from the Ground Truth. We didn't ask the AI to invent a racing line. We had a pro driver set a reference lap and coach a novice using VBOX video and telemetry to create a verified learning plan. This "Human-in-the-Loop" data became the foundation. The system wasn't guessing; it was scaling expert intuition. To deliver this insight at speed, the team architected a "Split-Brain" system: 🧠 Gemini Nano (The Reflexes): We pushed inference to the edge on Chrome. Nano handled safety-critical decisions—processing 10hz GPS telemetry with sub-50ms latency to determine track position—where speed is non-negotiable. ☁️ Gemini Flash 3.0 (The Strategist): We kept the heavy reasoning in the cloud. This layer acted as the "Race Engineer," comparing the novice’s performance against the expert's learning plan to identify strategic improvements. ⚡ Cloud Run & Vertex AI: The backbone. We used Cloud Run to ingest 100k+ data points without managing servers, while Vertex AI orchestrated the strategic inference pipeline. 🔗 Antigravity: This was the glue. It allowed the team to vibe code the integration between local edge interaction and cloud reasoning in a single, fluid workflow. It wasn't smooth sailing. We dealt with the "Reality of the Edge"—extreme vibration, heat, and intermittent connectivity. But that was the point. We proved that by combining the right models (Nano/Flash) with verified human expertise, you can build AI systems that aren't just smart, but trustworthy. Incredible work by the GDEs for architecting this: Austin Bennett, Hemanth HM, Jesse Nowlin, Jigyasa Grover, Lynn Langit, Margaret M., Rabimba Karanjai, Sebastian Gomez, Vikram Tiwari and the Product teams for giving us the tools to build it: Google DeepMind, Google Cloud, Logan Kilpatrick, Paige Bailey, Omar Sanseviero, Jetha Chan, Anshul Ramachandran, Naina Pasricha, Jason Mayes, Matt Thompson, Richard Seroter, Aja Hammerly Final shout out to Marty Fortier and Karen Acosta Parra for flawless logistics execution and Alvaro H. for technical support across the board! 📸 Margaret M. Lynn Langit Hemanth HM #TrustableAI #Antigravity #gemini #GeminiNano #GeminiFlash #CloudRun #webai #VertexAI #EdgeAI #Motorsport #google #googlecloud #gde #gdg Google Developers Group
-
-
-
-
-
+5
-
-
DataPrompt brings full type safety to your AI generations in Go ⚡
If you're using Go for AI, you probably want type safety on both sides of the prompt — the input and the output. DataPrompt lets you define typed inputs and get typed outputs. Your IDE autocompletes, your compiler catches mistakes. Type safety from input to output — that's how prompts should work in Go. Check out the docs to get started: https://lnkd.in/g8-87xdt
-
-
Google for Developers reposted this
It was an amazing experience and one incredible collaborative effort across domains! GDE++ 🤓 This is just the beginning. I’m really looking forward to more such experiments. Whether it’s connecting GPS devices to a laptop strapped into a race car and rendering real-time data on the UI along with nano, or talking to the CEO of the race track and getting expert pro driver assessments to improve agents, this was a fantastic AI field test. And there’s definitely more to come! Thank you, Ajeet Mirwanifor organizing this and giving us the thrill of being in a race car and taking agents to the race track! 🏎️ A big takeaway for me was how effortlessly AI can now interface with hardware, as well as discovering the exciting world of CAN bus (Controller Area Network). #AI #ML #WebAI
AI agents in a Jupyter notebook are safe. AI agents on a race track are a different story. For the High-Velocity AI Field Test, we issued a challenge to the Google Developer Experts (GDEs): Leverage the full Google stack to build a "Trustable AI" system that meets or exceeds the current state-of-the-art. The mission was to provide a driver with verifiable, real-time intelligence in a high-stakes environment where cloud latency isn't just a bug—it's a safety risk. The "Aha" Moment: The Human in the Loop - We quickly realized that trust doesn't come from the model alone; it comes from the Ground Truth. We didn't ask the AI to invent a racing line. We had a pro driver set a reference lap and coach a novice using VBOX video and telemetry to create a verified learning plan. This "Human-in-the-Loop" data became the foundation. The system wasn't guessing; it was scaling expert intuition. To deliver this insight at speed, the team architected a "Split-Brain" system: 🧠 Gemini Nano (The Reflexes): We pushed inference to the edge on Chrome. Nano handled safety-critical decisions—processing 10hz GPS telemetry with sub-50ms latency to determine track position—where speed is non-negotiable. ☁️ Gemini Flash 3.0 (The Strategist): We kept the heavy reasoning in the cloud. This layer acted as the "Race Engineer," comparing the novice’s performance against the expert's learning plan to identify strategic improvements. ⚡ Cloud Run & Vertex AI: The backbone. We used Cloud Run to ingest 100k+ data points without managing servers, while Vertex AI orchestrated the strategic inference pipeline. 🔗 Antigravity: This was the glue. It allowed the team to vibe code the integration between local edge interaction and cloud reasoning in a single, fluid workflow. It wasn't smooth sailing. We dealt with the "Reality of the Edge"—extreme vibration, heat, and intermittent connectivity. But that was the point. We proved that by combining the right models (Nano/Flash) with verified human expertise, you can build AI systems that aren't just smart, but trustworthy. Incredible work by the GDEs for architecting this: Austin Bennett, Hemanth HM, Jesse Nowlin, Jigyasa Grover, Lynn Langit, Margaret M., Rabimba Karanjai, Sebastian Gomez, Vikram Tiwari and the Product teams for giving us the tools to build it: Google DeepMind, Google Cloud, Logan Kilpatrick, Paige Bailey, Omar Sanseviero, Jetha Chan, Anshul Ramachandran, Naina Pasricha, Jason Mayes, Matt Thompson, Richard Seroter, Aja Hammerly Final shout out to Marty Fortier and Karen Acosta Parra for flawless logistics execution and Alvaro H. for technical support across the board! 📸 Margaret M. Lynn Langit Hemanth HM #TrustableAI #Antigravity #gemini #GeminiNano #GeminiFlash #CloudRun #webai #VertexAI #EdgeAI #Motorsport #google #googlecloud #gde #gdg Google Developers Group
-
-
-
-
-
+5
-