JSON in → App out: from tables and documents to graphs and agent memory
summary
Explains shifting from table/document data models to graph- and agent-memory-oriented models so frontend teams can build search engines, AI-first and production-ready features without deep backend work. Compares benefits and trade-offs of hybrid backend approaches in the age of ubiquitous vectorization for NLP pipelines and RAG. Demonstrates practical tooling: TypeScript 5+, Node.js, Vite, vibe-coding and RushDB.
AI-written summary · about
chapters (8)
- 0:00Welcome
- 7:09Traditional data storage problems
- 11:49MongoDB and schema flexibility
- 15:45Graph thinking for data
- 22:02HyperDB introduction
- 28:52HyperDB demo
- 39:20Graphs for frontend/backend
- 47:37Q&A
related talks
25mFrom Prompt to Production: Building Autonomous AI Agents in Node.js
48mGoing Beyond Prompt Engineering
24mEmpowering Vue Apps with AI
49mBuilding AI Enabled Web Apps
01h 17mGraph Database and Retrieval Augmented Generation
33mData.js
13mRAG Without the Hassle: Building AI-Powered Applications
35mDynamic, Safe Ops Using AI to Generate Trusted Operations from Text Prompts