Designing AI/ML apps vs traditional web , client/server, mobile apps - Things to keep in mind

Google Cloud Compute Infrastructure
1

Is there any good documentation to designing AI/ML apps on GCP? I see that architecting AI/ML apps is way different from designing traditional web, mobile apps. Some of the considerations are

  1. RAG vs non-RAG
  2. Vector DB or traditional relational DB
  3. Choosing the right model /ML framework like Langchain for our use case
  4. LLM vs non LLM (not sure if this is correct)
2

Hi @dheerajpanyam,

Welcome to Google Cloud Community!

Yes, you’re right! Designing AI/ML applications requires specific considerations that differ from traditional web and mobile app development. It demands a distinct mindset, prioritizing data preparation, computational efficiency, and model selection.

To design AI/ML apps on Google Cloud Platform, you may want to review the following documentation for in-depth guidance on getting started, training models, and deploying them effectively:

  • Google Cloud AI and ML documentation - it offers an introduction to machine learning on Vertex AI and facilitates workflows in data science, data engineering, and ML engineering.
  • Architecture Framework - Explain the best practices and guidelines for designing, building, and managing AI/ML tasks in Google Cloud.

I hope the above information is helpful.

1 Like