I architect systems at the boundary of code and inquiry — from causal AI research platforms to immersive web experiences. Development is my form of publication. Every system I build is treated with the rigor of a research paper and the precision of a surgical instrument — functional, communicative, and hard-to-vary by design.
🔬 wuweism.com — MASA: a trust-first AI architecture for scientific discovery, causal governance, and auditable reasoning.
- 🧪 Crucible — MASA research workbench. The production surface for closed-loop scientific analysis.
- 🛡️ academic-integrity-agent — AI verification agent with burstiness analysis, perplexity tracking, and human variance preservation.
- ⚡ voltagent-crucible-client — VoltAgent client bridge for Crucible
- 🔌 masa-orchestrator-mcp — MCP orchestrator for MASA multi-model pipelines
- 🧬 hiv-entropic-vise — Entropic analysis framework for HIV research
My primary work is MASA — a causal AI research platform built on Pearl's do-calculus, designed to close the loop between hypothesis, intervention, and counterfactual reasoning.
- Closed-loop AI scientist — Hypothesis generation → causal evidence extraction → claim falsification → epistemic memory
- Causal Engine — Graph surgery, forward substitution, structural counterfactuals. If the LLM disappeared, the core computation would still run.
- Sovereign Memory — RAG architecture that tracks confidence, not just recall.
- Read the white paper →
Pearl's do-calculus. Popper's falsifiability. Deutsch's criterion that good explanations are hard to vary. Maltz's cybernetic self-image. These are not influences — they are load-bearing walls.
Causal inference · Theoretical neuroscience · Philosophy of science · Higher-order cognition
Photography is how I train my eye. The same discipline that produces a decisive frame produces a decisive interface.
为学日益，为道日损
In pursuit of knowledge, every day something is added. In pursuit of the Way, every day something is dropped.
— Tao Te Ching, Ch. 48