Refactor Architecture Diagram to Align with New Data ML → Gen AI Core Flow

[SwathiKrish97]

Background

Our Data Engineering and AI processing approach has been updated to a real-time,
GenAI-driven flow.

This task is to redesign the architecture diagram so it accurately reflects
the new core runtime flow and removes legacy analytics components from the
primary execution path.

Objective

Update the architecture diagram so it:

• Accurately represents the new Data ML → Gen AI runtime flow
• Minimizes database storage usage
• Removes legacy analytics components from the core execution path
• Clearly shows API-based invocation from the frontend via API Gateway

Approved Core Runtime Flow

1. Frontend invokes an API exposed via API Gateway.
2. API Gateway routes the request to the Data ML microservice.
   • Request includes beneficiary information, help category, and optional organization hints.
3. Data ML calls the Gen AI microservice.
4. Gen AI:
   • Classifies the help category
   • Fills missing fields
   • Generates organization information:
     • Nonprofits → enriched using public data sources (e.g., ProPublica)
     • For-profits → structured data generated via LLM
5. Data ML checks the database:
   • If the organization exists → fetch and enrich data
   • If the organization does not exist → do not store yet
6. Only when a final beneficiary–organization match is confirmed:
   • Insert the organization into the database
7. Data ML returns one combined JSON response back to the Frontend.

Scope

• Architecture diagram update only
• No implementation or code changes in this task