How Kakao Built an AI-Powered Internal Service Desk with Qdrant - Qdrant

What this post added

This post details Kakao's implementation of an AI-powered internal service desk using Qdrant as a RAG system. It highlights the use of Qdrant's hybrid search (dense and sparse vectors with RRF) to handle both semantic and exact-match queries across technical documentation and historical Q&A data. The implementation uses Named Vectors, scalar quantization (INT8) for memory optimization, and an asynchronous Python client for high concurrency. Qdrant is self-hosted on Kubernetes with a replication-based setup for high availability. The system supports approximately 1 million vectors and has reduced end-to-end response times to under 30 seconds.