10 Google AI Tools for GenAI Developers You Need to Know
When it comes to open source technology, Google is topping the charts on the Open Source Contribution Index. But beyond their open-source contributions, Google also offers an impressive array of tools that many developers may not even know about. In this blog post, we’ll shine a spotlight on my selection of 10 tools for AI Developers. Whether you’re a seasoned coder or just starting out, these tools are definitely worth exploring and incorporating into your workflow. Let’s dive in!
1. Google AI Studio: https://ai.google.dev/
Google AI Studio is a browser-based IDE for prototyping with generative models from Google. This includes both best in class Gemini models as well as Gemma, which represents a family of lightweight, state-of-the art open models built from the same research and technology used to create the Gemini models. Ideal environment for validating business prototypes experimenting with prompt engineering, fine-tuning...
However, as your generative AI solutions mature, you may need a platform for building and deploying generative AI applications and solutions end to end. And this is when you may need Google Cloud. Google Cloud provides a comprehensive ecosystem of tools to enable developers to harness the power of generative AI, from the initial stages of app development to app deployment, app hosting, and managing complex data at scale. You will want to transition to Vertex AI platform which offers a suite of MLOps tools that streamline usage, deployment, and monitoring of AI models for efficiency and reliability. Google Cloud will also give you native integrations with databases, DevOps tools, logging, monitoring, and IAM meaning it will provide a holistic approach to managing the entire generative AI lifecycle.
2. IDX: https://idx.google.com/
Project IDX is an AI-assisted, entirely web-based workspace for full-stack, multiplatform app development in the cloud.
What is there fore you? You can build and deploy applications in various tech stacks without the tedium of setting up development environments for yourself and others on your team.
IDX uses Google Cloud to keep your development environment reliable, safe, and fully customizable, but you don’t need Google Cloud account to use it! It utilizes the same environment as Google Cloud Workstations which is a cloud service providing fully managed development environments built to meet the needs of security-sensitive enterprises. Cloud Workstations allow you to develop and run code inside your private network and in your staging environment, so you don’t need to emulate your services. You can also enforce “no source code on local devices” policies and bring the same security mechanisms used for production workloads to your development environments, such as VPC Service Controls (VPC SC), private ingress/egress, Cloud Audit Logs, and granular IAM controls.
3. Colabs: https://colab.research.google.com/
Colab is a web-based Notebook service that requires no setup to use and provides free access to computing resources, including GPUs and TPUs. Colab is especially well suited to machine learning, data science, and education (e.g. to lean python, JAX, Tensorflow, Pytorch…).
If you need:
- More compute units
- Faster GPUs
- More memory
- Terminal access
- Background execution
then these will be available from Google Workspace:
or Google Cloud Vertex AI Colabs:
4. Visual Blocks for ML: https://visualblocks.withgoogle.com/#/
This one has such a huge potential! Visual Blocks for ML is a Google visual programming framework that lets you create ML pipelines using a no-code graph editor. You can quickly prototype workflows by connecting drag-and-drop ML components for vision, text, audio and video, including models, user inputs, processors, and visualizations. You can also code your custom ML components. What is extremely helpful is the fact that you can run your pipelines in real-time in your browser and preview results at each step. This streamlines the development and debugging process, allowing you to quickly identify issues and iterate on your pipelines without the need for constant code re-execution and waiting for results.
If you would like to use this from your laptop or from Google Cloud then Visual Blocks can be installed into any Jupyter Notebook, including Colabs available from Workspace or Vertex AI:
5. LLM Comparator: https://pair-code.github.io/llm-comparator/
Code: https://github.com/PAIR-code/llm-comparator
LLM Comparator is an interactive visualization tool with a python library, for analyzing side-by-side LLM evaluation results. It is designed to help people qualitatively analyze how responses from two models differ at example- and slice-levels. Users can interactively discover insights like “Model A’s responses are better than B’s on email rewriting tasks because Model A tends to generate bulleted lists more often.”
Why side-by-side LLM evaluation? The notion that your AI model is always superior is a common misconception fueled by the rapid advancements and marketing hype surrounding artificial intelligence. In reality, the AI landscape is vast and diverse, with a multitude of models excelling in different areas. The only way to find out which model works best for you is to measure its performance on your use case and then consider other factors like latency, cost, ease of use, stability, availability ….. But the very first step is to measure if you assumptions hold true.
In fact when you think about it, all these shows that it’s not just about the model itself; it’s about the platform that empowers you to create trustworthy AI applications and guides you in making informed choices. Vertex AI exemplifies this, offering a comprehensive suite of tools and resources to build, deploy, and manage AI models effectively. It’s the platform that truly unlocks the potential of AI, not just the model alone.
One of the tools it offers to make informed decisions about the model is automated side by side (AutoSxS) comparison of distinct models on your inputs.
To understand how it works, imagine you and your friend are playing a drawing game. You both get the same instructions, like “draw a cat wearing a hat,” and you both draw your own versions. Now, imagine there’s a third person, like a teacher or a parent, who looks at both drawings and decides which one is better.
In the world of AI, this is kind of like how AutoSxS works. There are two AI models, let’s call them Model A and Model B. They both get the same instructions, or “prompts”. Each model writes its own response, just like you and your friend drew your own cats.
Then, there’s another AI model, called the “autorater”. This is like the teacher or parent in our drawing game. The autorater looks at both stories and decides which one is better. It does this by using a set of rules, or “criteria,” to judge things like how well the stories follow the instructions, how interesting they are, and how well they’re written:
The autorater then tells us which model did a better job overall, and it can even explain why it chose one story over the other. It also tells us how confident it is in its decision, just like a teacher might say “I’m very sure your drawing is better” or “It’s a close call, but I think your friend’s drawing is slightly better.”
This is how AutoSxS helps us figure out which AI model is better at a particular task, like writing stories or answering questions. It’s like a fair and objective judge for AI models.
Once the rating is done you can visit Vertex AI console and see a list of your input questions, and for each question, you’ll know which model’s answer the autorater preferred. The autorater explains why it chose that answer and how confident it feels about its judgment. The win rate tells you which model overall performed better across all your questions.
Remember, even though there’s only one winner per question, both models might give correct answers. It’s just that one answer might be better based on the criteria the autorater used, like being more informative or easier to understand. But remember this universal truth: you can not improve things that can’t be measured so make data driven decisions. AutoSxS is a very powerful tool in your toolkit.
So where is LLM Comparator needed? LLM Comparator tool computes aggregated statistics that help to understand when, how and why the winner is better.
6. Firebase: https://firebase.google.com/
Firebase is a comprehensive app development platform that is globally used by millions of mobile applications and websites. Firebase handles many backend tasks, such as authentication, real-time database, cloud storage, and hosting, enabling developers to focus on creating compelling user experiences rather than building and managing complex infrastructure.
For generative AI developers, Firebase becomes even more valuable when combined with tools like Genkit or native integration with Vertex AI or Google AI Studio.
Few words about Genkit. Genkit is a framework designed to simplify the integration of AI capabilities into applications. It is available as open source libraries for Node.js and Go, plus developer tools for testing and debugging
These tools help you:
- Experiment: Test and refine your AI functions, prompts, and queries.
- Debug: Find and fix issues with detailed execution traces.
- Evaluate: Assess generated results across multiple test cases.
7. Mesop: https://google.github.io/mesop/ (showcase: https://wwwillchen-mesop-showcase.hf.space/)
Mesop is a Python-based UI framework that allows you to rapidly build web apps like demos and internal apps:
- It is Open-Source,
- it is very easy to get started with pre-built components,
- it enables you to code in idiomatic Python,
- it supports hot-reloads,
- its components are basically Python Functions,
- it is fast and built on Angular.
- It is my default choice whenever I need to build something that requires user interface.
What is quite interesting is that it can also be used in interactive mode from Colab:
8. OneTwo: https://github.com/google-deepmind/onetwo
OneTwo is a Python library designed to simplify interactions with large (language and multimodal) foundation models, primarily aimed at researchers in prompting and prompting strategies.
Some properties of OneTwo that are particularly impactful for researcher productivity include the following:
- Model-agnostic: Provides a uniform API to access different models that can easily be swapped and compared.
- Flexible: Supports implementation of arbitrarily complex computation graphs involving combinations of sequential and parallel operations, including interleaving of calls to foundation models and to other tools.
- Efficient: Automatically optimizes request batching and other details of model server interactions under-the-hood for maximizing throughput, while allowing prompting strategies to be implemented straightforwardly, as if they were dealing with just single requests.
- Reproducible: Automatically caches requests/replies for easy stop-and-go or replay of experiments.
You can learn more from my articles:
- From Zero to Hero: Building LLM Chains and Agents with Google DeepMind’s OneTwo OpenSource Toolkit:
- OneTwo and Vertex AI Reasoning Engine: exploring advanced AI agent development on Google Cloud:
9. Breadboard: https://breadboard-ai.web.app
Code: https://github.com/breadboard-ai/breadboard
Breadboard is a library for prototyping generative AI applications. The best way to get started with Breadboard is to use the Visual Editor an follow this guide.
You can also use TypeScript API. Breadboard can be deployed on your laptop or on Google Cloud.
10. MediaPipe: https://mediapipe-studio.webapps.google.com/home
Code: https://github.com/google-ai-edge/mediapipe
MediaPipe Solutions provides a suite of libraries and tools to quickly apply artificial intelligence (AI) and machine learning (ML) techniques in your applications. You can plug these solutions into your applications immediately, customize them to your needs, and use them across multiple development platforms. MediaPipe Solutions are available across multiple platforms. Each solution includes one or more models, and you can customize models for some solutions as well.
You can try it online: https://mediapipe-studio.webapps.google.com/home
The following list shows what solutions are available for each supported platform and if you can use Model Maker to customize the model:
Summary:
In this rapidly evolving field of generative AI, Google is paving the way with an impressive suite of tools that streamline the development process, enhance model evaluation, and simplify the integration of AI capabilities into applications. From prototyping to deployment, these tools offer a comprehensive solution for building cutting-edge GenAI solutions.
Whether you’re a seasoned AI developer or just starting your journey, exploring these tools can significantly boost your productivity and unlock new creative possibilities. I encourage you to dive deeper into the tools that pique your interest, experiment with their features, and discover how they can improve your GenAI development workflow.
And don’t hesitate to share your experiences and insights with the community! Let’s collaborate, learn from each other, and push the boundaries of what’s possible with generative AI together.
Remember, the future of AI is in your hands. Embrace these tools, explore their potential, and let your creativity soar!
This article is authored by Lukasz Olejniczak — Customer Engineer at Google Cloud. The views expressed are those of the authors and don’t necessarily reflect those of Google.
Please clap for this article if you enjoyed reading it. For more about google cloud, data science, data engineering, and AI/ML follow me on LinkedIn.
