Technical resources and projects centered on ai.
A battle-tested blueprint for building AI agents that interact with local and cloud data via the Model Context Protocol.
A technical comparison of Retrieval-Augmented Generation (RAG) vs Cache-Augmented Generation (CAG) for personal and enterprise knowledge bases.
Designed and built a full-stack AI fashion platform on a Bun monorepo. Turns flat lay product photos into photorealistic campaign visuals through prompt-engineered Gemini vision models.
A high-performance Model Context Protocol (MCP) server giving AI tools direct, graph-aware access to an Obsidian vault. Hybrid SQL/vector search with 50ms tool response.
How I turned Obsidian into a structured context layer for AI systems using MCP, and why context architecture matters more than model scale.
RAG is a search bar. Agents need navigation. 3 patterns: Tool-first design, Schema-aware retrieval, and Stateful reflection.
Demos are easy. Production is hard. Lessons from Anaqio: Prompt-as-code, Generative Error Handling, and Boring State Machines.
Vector search is a bottleneck. I moved to Context-Augmented Generation (CAG). Large windows + Prompt Caching = Superior Reasoning.