What to Build

Your mission is to build an innovative serverless application deployed on Google Cloud Run. Your project should fit into one of the following categories:

Categories

AI Studio Category

• You must use AI Studio to generate a portion of your application
• Your application must be deployed to Cloud Run. You can build and deploy your workload directly on to Cloud Run, or use the Deploy to Run button in AI Studio
• Challenge: Take an idea and make it into code. Use Google AI Studio to vibe-code your concept and deploy the app directly to Cloud Run.
• Details: Your project could be a web service, a data processing pipeline, a backend for a mobile app, or any other creative serverless solution.

AI Agents Category

• Your agent must be built with Googles Agent Development Kit (ADK)
• Your agents must be deployed to Cloud Run
• Challenge: Build a multi-agent application and deploy it to Cloud Run.
• Details: Create an agent-based application using the Google Agent Development Kit (ADK). Your solution should consist of at least two AI agents that communicate to complete a workflow, solving a real-world problem or improving a process.

GPU Category

• You must utilize the provided Nvidia L4 GPUs on Cloud Run, either as a Service, Job or Worker Pool
• You should use the europe-west1 or europe-west4 region for your GPU workloads
• Challenge: Harness the power of GPUs on Cloud Run to run performant AI and ML models.
• Details: Your project must deploy and run an open-source model (like Gemma) on a Cloud Run service configured with the provided NVIDIA L4 GPU.

General Requirements (All Categories)

Your project must be deployed on Google Cloud Run. Feel free to use one (or more) of the three resource types:

1. Service: For handling HTTP requests, events, and functions.
2. Job: For parallelizable tasks that run to completion.
3. Worker pool: For pull-based workloads like Kafka or Pub/Sub consumers.

Allowed Technologies (Optional)

While Cloud Run is the star, we encourage you to enhance your solutions by integrating with other technologies in the Google Cloud ecosystem:

1. Foundation Models: We highly encourage the use of Gemini, Gemma, Imagen, and Veo.
2. CLI-based AI Assistance: Use Gemini CLI to accelerate your development workflow.
3. Google Cloud Services: Integrate with Cloud Storage, BigQuery, Firestore, and more!
4. IoT & Mobile devices: Your Cloud Run application can power more than just web services!

What to Submit

• Text Description: A summary of your project's features, functionality, and the tech you used.
• Demonstration Video: A video (around 3 minutes max) that walks through your project.
• URL to your Public Code Repository: Let us see how you built it!
• Architecture Diagram: A visual representation of your project's components and data flow.
• Try it Out Link: A URL to your hosted project (if applicable), highly encouraged.
• **AI Studio Category: Provide a link to your prompts in AI Studio by sharing the link to your app using the Share App functionality in AI Studio

Optional Google Cloud Contributions: (a maximum of 0.4 points will be added) 

1. The use of Google AI models: Including but not limited to models like Gemini, Gemma, Veo, etc
2. Leveraging more Cloud Run Services: points will be added if the project runs multiple services, need a front & back end 

Optional Developer Contributions: (a maximum of A maximum of 0.4 points will be added per contribution type)

1. Publish a piece of content (blog, podcast, video): Cover how the project was built using Cloud Run on any public platform (e.g., medium.com, dev.to, Youtube, etc.). The content must be public (not unlisted). You must include language that says you created the piece of content for the purposes of entering this hackathon.
2. Publish a social media post: Highlight or promote your project on a social media post on X, LinkedIn, Instagram, or Facebook. Include the hashtag #CloudRunHackathon.

See Full Rules for Details. 

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.