PJPraise Jamesinactiandev.hashnode.dev·Jul 27 · 11 min readBuild a local RAG pipeline on declassified UFO files with VectorAI DB and OllamaProduction document collections rarely look like benchmark datasets. Some PDFs contain searchable text. Others are scanned pages that need OCR. Metadata varies across sources, and retrieval pipelines 00
PJPraise Jamesinactiandev.hashnode.dev·Jul 27 · 15 min readBuild a Local Document Intelligence Agent with LlamaIndex and VectorAI DBMost LlamaIndex vector store integrations assume an outbound connection. Pineone, Weaviate, Qdrant Cloud, and the rest of the officially supported list all require a network call to an external servic00
PJPraise Jamesinactiandev.hashnode.dev·Jul 15 · 18 min readMulti-Collection Vector Search for Clinical DataYou are testing your clinical search system and ask it to find patients with lab profiles similar to a reference case for chronic kidney disease. The results come back with full discharge summaries an00
PJPraise Jamesinactiandev.hashnode.dev·Jul 15 · 14 min readHow to Build Persistent Agent Memory Across SessionsYour agent can follow a rule perfectly at the start of a conversation and still miss it later. Maybe the rule is simple: don't share user contact details, don't book meetings outside business hours, o00
PJPraise Jamesinactiandev.hashnode.dev·Jul 15 · 16 min readRunning Gemma 2B on Edge Hardware with Actian VectorAI DBToday, running a powerful language model entirely on edge hardware is no longer the hard part. The problem is making it useful once it is there. A developer can deploy Gemma 2B on NVIDIA Jetson Orin N00
