QuantVault Second Brain
A 10,743-entry Obsidian knowledge base spanning eight domains. Not a notes folder — a structured, cross-linked knowledge graph with a consistent authoring methodology and a retrieval layer wired directly into Claude.
Reading about quantitative finance produces the illusion of understanding and very little retention. Notes accumulate, nothing links, and six months later you cannot find the one paragraph that mattered. The problem is not note-taking, it is architecture: without structure and retrieval, a pile of notes is write-only.
The parts that carry the weight.
Eight domains, one graph
Strategies, mathematics, market microstructure, risk, infrastructure, market history, behavioral finance, and careers. Every entry lives in one domain and links across many, so the structure is a graph rather than a filing cabinet.
Wave-based authoring
Entries are written in deliberate waves against a consistent methodology rather than accumulated ad hoc, which is what keeps 10,743 of them coherent instead of merely numerous.
A tagged taxonomy
One consistent tag vocabulary across the whole base, so retrieval can rely on the structure being the same everywhere rather than on remembering how a given note was filed.
Wired to Claude
A retrieval layer connects the vault directly into Claude, so it works as a live reasoning surface rather than a static archive. The value is not that the notes exist, it is that they can be reasoned over.
Why it's built this way.
- 01
Designed as a knowledge graph rather than a notes folder from the beginning. Retrofitting structure onto an unstructured pile is much harder than authoring into one.
- 02
RAG-ready structure is a design constraint, not an afterthought — entries are written at a size and specificity that retrieval can actually use.
- 03
Domain boundaries are deliberately coarse. Fine-grained folders push you into filing decisions that links should be making instead.