Product-search assistant
A chatbot that finds products from a plain question, using vector search and a graph of how products relate to each other.
- Role
- Software engineer
The problem
People describe what they need in their own words, and products relate to each other in ways a keyword search does not see.
What I built
- A product-search chatbot, built on a RabbitMQ ingestion pipeline.
- Answers that combine Qdrant vector search with product relationships modeled in Neo4j.
Technical details
- Product data flows in through a RabbitMQ ingestion pipeline.
- Qdrant handles vector search; Neo4j holds the relationships between products.
Stack
- RabbitMQ
- Qdrant
- Neo4j