Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Link Library
A technical walkthrough of building LinkLibrary, covering data modeling, RAG implementation, AI agent experiments, and architectural tradeoffs using FastAPI and Postgres.
LinkLibrary started as a personal frustration: saving links everywhere (Notion, Notes, Slack, browser bookmarks) and never finding them again. In this talk, I’ll walk through how LinkLibrary was actually built, from early scrappy experiments to a more AI-native system.
I’ll show:
• How we designed the core data model for links, collections, and social signals
• How we approached semantic chunking, embeddings, and retrieval (what worked and what didn’t)
• Early attempts at AI agents for tagging, summarization, and discovery
• Technical tradeoffs across FastAPI, Supabase/Postgres (pgvector), and React
• Lessons learned from shipping v1, demoing to real users, and iterating in public
This is a behind-the-scenes, technical walkthrough focused on architecture, decisions etc