The overwhelmed person’s guide to Google Cloud: week of July 18
Richard Seroter
Chief Evangelist, Google Cloud
A weekly curation of the most helpful blogs, exciting new features, and useful events coming out of Google Cloud.
The content in this blog post was originally published last week as a members-only email to the Google Cloud Innovators community. To get this content directly in your inbox (not to mention lots of other benefits), sign up to be an Innovator today.
New and shiny
Three new things to know this week
- Control Gemini model output with JSON schemas. With Gemini 1.5 Pro, we now offer the ability to control the generated output format from the model. Your app may need predictable response formats from the LLM calls, and Vertex AI makes this possible. See details and examples here.
- Share streaming data with other internal teams, or outside subscribers. Our Analytics Hub gives you a way to monetize your datasets and sell access to external clients. We just added the option to share Pub/Sub Topics too, meaning that you can offer public or private access to data subscribers. Very cool.
- Generate SLSA v1.0 build provenance in Cloud Build. Contribute to a strong security posture by producing a build provenance that explains your software artifacts and sources. Cloud Build does this for you automatically. Learn more about how to build, review, and verify the provenance.
Watch this
Write even better prompts for Gemini Code Assist. Watch as Luis makes us smarter at writing effective prompts that help Gemini Code Assist generate the code we’re looking for.
Community cuts
Every week I round up some of my favorite links from builders around the Google Cloud-iverse. Want to see your blog or video in the next issue? Drop Richard a line!
- Scientific workflows with Nextflow on Google Cloud. Lynn has an excellent post that digs into the architecture she constructed. She uses Google Batch, Cloud Shell, GCE, Artifact Registry, and more.
- Building AI agents with structured data sources. If you want to ask questions of your CRM data via a generative AI agent, what are your options? Simon shows off a good one in this post, as he explores Vertex AI Agent Builder interfaces to store data and build a chat interface.
- Handle long-running services with Cloud Run and Firestore. This post from Dale features a good discussion of how to handle long-running requests. He then demonstrates a solution made up of Cloud Storage, Cloud Run, and Firestore with an Angular frontend.
Learn and grow
- Deploy your Go apps to Cloud Run directly from source control. This codelab helps you learn how to set up automatic deployments of your Go code. You can follow along fairly easily, and it all fits within the free tier of each service. Check out the equivalent codelab for Node.js apps.
- How do you use the BigQuery data canvas? Here’s how. This AI-driven analytics surface is now generally available. Read this post for a useful dive into what it takes to find, query, and visualize your BigQuery data. And don’t miss this companion post on how to write good prompts for the data canvas.
- You want advanced CI, I’ll give you advanced CI. Darren continues his look at optimized continuous integration with this post that features the Bazel build tool, amongst other things. He looks at how to do remote builds on GKE and parallel builds steps in Cloud Build.
- Use Vertex AI Search to ground your LLMs. You’ve got private data, and an app that wants contextual LLM results? Mete shows us how to use a PDF for grounding in Vertex AI Search, and how the results differ from vanilla LLM calls. He’s also got a related post about doing this with the LlamaIndex framework.
One more thing
Alok talks about AI in sports on the Wharton Moneyball podcast. Various sports teams are using AI to improve their analysis, and Alok joined this Wharton School podcast to share his insights.
Become an Innovator to stay up-to-date on the latest news, product updates, events, and learning opportunities with Google Cloud.
