Skip to main content

Topic 8: Cloud AI Service Platforms

⏱️ Estimated time: 3-4 days

Cloud providers offer comprehensive AI service platforms. In this topic, you'll learn about model options, deployment strategies, and operational features. Understanding these platforms helps you make informed architectural decisions when integrating AI into applications.

📚 Learning Path​

  • Study - Be sure to read additional relevant articles in respective documentation sections as necessary

Platform Overview​

Azure | AWS | GCP

About Models​

Azure | AWS | GCP

Deployment​

Azure | AWS | GCP

Safety & Guardrails​

Azure | AWS | GCP

Platform Playgrounds​

Azure | AWS | GCP

🧪 Test Your Knowledge​

Once you've reviewed your chosen platform, test your understanding with these prompts:

  1. What are some models that your chosen platform offers? What are their key differences?
  2. When would you choose a smaller, faster model vs. a larger, more capable model for your journal analysis use case?
  3. How do guardrails and content filters work in your platform, and when would you use them?
  4. What are the cost implications of your model choice and deployment strategy?
  5. Describe the deployment options available in your platform. How would you deploy a model for your journal application?
  6. How would you test different models and prompts before deploying to production?

🛠️ Hands-on Task​

Explore your cloud AI platform and deploy a model:

  • Compare at least 2 different models in your platform's playground tool
  • Test different prompts and observe how the models respond (quality, speed, cost differences)
  • Analyze the available deployment options and their trade-offs
  • Understand which model and deployment approach you'd choose for your journal application and why
  • Deploy a model that you will utilize for the capstone

✅ Topic Checklist​

Before moving on, make sure you have:

  • Understood what your platform offers
  • Explored the available models on your chosen platform
  • Tested multiple models or prompts in your platform's tools
  • Understood deployment options and their differences
  • Reviewed safety/guardrail features
  • Made an informed decision about which model to use for the capstone