Building Advanced LLM Applications with Python and Milvus: A Scalable Vector DB Solution
summary
Shows a working RAG pipeline that uses Milvus as the vector store, LangChain for orchestration, and LLaMA/ollama models for generation, with Milvus local-to-cloud connector letting the same Jupyter code run against local, Docker or cloud instances. Demos include all‑text PDFs embedded with all-MiniLM embeddings, a street‑camera pipeline producing two vectors per record (image via ResNet34 and text/metadata), YOLO object detections, Raspberry Pi edge inference, thousands of points indexed and queried in milliseconds. RAG with LangChain is implemented end-to-end and runs successfully.
AI-written summary · about
chapters (15)
- 0:00Introduction and setup
- 1:04Unstructured data and vectors
- 3:14Vector database options
- 4:42Retrieval Augmented Generation
- 7:56Demo environment setup
- 10:29Loading and processing documents
- 13:00Building the RAG application
- 14:48First demo walkthrough
- 17:17Loading unstructured data
- 18:22Street camera data processing
- 20:26Live plane tracking demo
- 21:10Metadata and vector fields
- 22:19Real-time data search
- 23:47Air quality data demo
- 24:08Resources and wrap-up
related talks
31mBuild an Agentic RAG system using Langchain, Ollama and Milvus
37mMilvus: Scaling Vector Data Solutions with Gen AI- 37mMilvus: Scaling Vector Data Solutions with Gen AI
01h 19mBuilding an LLM on your Laptop to...
51mUnleashing the Potential of Cloud-Native Vector Databases
36mMilvus: Scaling Vector Data Solutions with Gen AI
30mMilvus Vector Database: Unlocking the Future of Open Source Vector Databases
26mThe Rise of Vector Databases - Lessons from the Milvus Community