For anyone tracking the evolution of AI, reading articles is one thing—but actually getting your hands dirty with real development is where the magic happens. I recently earned the Build with Gemini Skill Badge from Google Cloud on Google Cloud Skills Boost. If you are looking for a practical, zero-cost way to transition from basic prompt engineering to building production-ready apps, this is it.
I have been reading about Gemini for a while, but this was my first hands-on experience actually building applications with it on Vertex AI. It is part of Google's GenAI Exchange program on Google Cloud Skills Boost.What is this badge about?
This badge is not just about watching videos. The path consists of four targeted, hands-on labs that demystify the process:- AI Image Recognition App: Building a framework to upload images and ask contextual, multimodal questions.
- AI Image Generator App: Constructing text-to-image endpoints to generate stunning assets from descriptive prompts.
- Streaming Chat Integration: Setting up multi-turn conversations and seeing how structural adjustments dramatically change output.
- The Challenge Lab: A true standalone test where you must synthesise Gemini and Imagen into one cohesive, multi-modal cloud application without step-by-step instructions.
The key takeaway for me was how easy Vertex AI SDK makes it. You can use the same SDK for text, image understanding, and image generation.
The idea is simple: FloraCare helps plant enthusiasts choose ideal indoor and outdoor plants, troubleshoot care needs, manage greenhouse inventory, compute fertilizer dosages, and visualize plants using generative multimodal tools and rich A2UI cards. You upload a photo of your plant that looks sick, and the agent identifies the issue and gives you care tips. It can also generate an image of a healthy plant for reference.
I used Gemini 2.0 Flash for its multimodal capability and Imagen for generation. The whole app is built on Vertex AI with a Streamlit UI. By pairing Gemini 2.0 Flash for rapid multimodal analysis with Imagen for asset generation, the application showcases how enterprise-grade AI models can easily solve everyday problems.
I have open-sourced it here for anyone to try:
GitHub: https://github.com/anjuprasannan/buildwithgemini-floracare-agent.git
This badge was a great starting point to go from prompting to building. If you are interested in GenAI, I would highly recommend taking this Skill Badge on Google Cloud Skills Boost - it takes about 4 hours.
Badge: Build with Gemini
Issued by: Google Cloud
#BuildWithGemini #GenAIExchange #GoogleCloud #VertexAI
From Badge to Project: FloraCare Agent
After finishing the labs, I wanted to build something of my own. So I created FloraCare Agent - a small AI agent for plant care.The idea is simple: FloraCare helps plant enthusiasts choose ideal indoor and outdoor plants, troubleshoot care needs, manage greenhouse inventory, compute fertilizer dosages, and visualize plants using generative multimodal tools and rich A2UI cards. You upload a photo of your plant that looks sick, and the agent identifies the issue and gives you care tips. It can also generate an image of a healthy plant for reference.
I used Gemini 2.0 Flash for its multimodal capability and Imagen for generation. The whole app is built on Vertex AI with a Streamlit UI. By pairing Gemini 2.0 Flash for rapid multimodal analysis with Imagen for asset generation, the application showcases how enterprise-grade AI models can easily solve everyday problems.
I have open-sourced it here for anyone to try:
GitHub: https://github.com/anjuprasannan/buildwithgemini-floracare-agent.git
This badge was a great starting point to go from prompting to building. If you are interested in GenAI, I would highly recommend taking this Skill Badge on Google Cloud Skills Boost - it takes about 4 hours.
Badge: Build with Gemini
Issued by: Google Cloud
#BuildWithGemini #GenAIExchange #GoogleCloud #VertexAI
No comments:
Post a Comment