Case study: LLM wiki à la Karpathy – context engineering in practice
Context engineering as a concept: a personal, growing wiki as a local context source for LLM queries – structured instead of naive vector search.
Context engineering as a concept – the core positioning.
Tech stack: TypeScript · PostgreSQL · Qwen3 · Ollama
Problem
Notes, learnings and research were scattered across loose files. To feed an LLM meaningfully with my own knowledge, a structured, continuously growing source was missing – instead of isolated snippets.
My role
Concept and implementation single-handedly – data model, backend and local LLM integration.
Solution
A personal wiki following the Karpathy pattern: knowledge is accumulated continuously and stored in a structured way in PostgreSQL, TypeScript as the end-to-end stack, Qwen3 via Ollama for local LLM features. The focus is context engineering – the wiki feeds precisely the right context into the model instead of relying on naive vector search.
Result
A personal, growing knowledge base that serves as a local context source for LLM queries – fully offline, no cloud.
Takeaways
Context engineering in practice: structured, accumulated notes beat naive vector search. The value comes from consistent upkeep – a wiki is only as good as the discipline behind it.