[ossamartinaaron09-ui]

Summary

Adds an end-to-end working example demonstrating how Dozer real-time CDC combines with ChromaDB vector search and LangChain to build hyper-personalized financial chatbots (#1690), following the Dozer LLM Hyper-Personalization Article.

Implementation Details

• Dozer Streaming Pipeline (dozer-config.yaml): Sources live PostgreSQL customer profiles and transaction records, serving aggregated real-time spending habits at /customer_profiles.
• Vector Knowledge Base (rag_pipeline.py): Indexes banking credit card products into ChromaDB with metadata filtering (minimum credit score, reward categories).
• LangChain Integration: Constructs hyper-personalized contextual prompts synthesizing Dozer customer spending metrics with vector search retrieval.
• Docker Compose Stack (docker-compose.yaml): Full reproducible setup with PostgreSQL CDC, Dozer Engine, and Python application.
• Unit Test Coverage (test_rag_pipeline.py): 100% test pass validating ingestion, retrieval, and prompt generation.

/claim #1690

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}