Neuroscience → AI Systems · Behavioral Finance
I study how people make decisions at the neural level — reward, motivation, cognitive bias, risk tolerance — and then I build systems that account for those realities. Production-deployed, not notebooks.
Live systems with real stakes. Each links to source.
Neural content intelligence — scores short-form video by predicted brain response, trained on fMRI data, shipped with a pre-registered validation protocol so the claim can be proven wrong.
Real-time ETF sector-rotation analyzer. Ingests public SSGA fund-flow data + price history, detects regime/rotation signals, broadcasts over WebSocket — and renders the live signal on a physical e-ink panel. Decision-support only.
Self-hosted personal memory & RAG. A LanceDB vault, an MCP query API, and a daemon fleet that ingests and maintains a knowledge graph ambiently — with a 3-tier watchdog that self-heals before it pages me.
AI-native CRM + PM intelligence. A unified sales pipeline and communication- extraction platform that turns scattered signal into structured next-steps — live in production.
An AI that argues with itself so you don't have to. Three agents with competing cognitive-bias profiles — Entrepreneur, Quant, Auditor — debate every strategic input, then route the work into action automatically.
Sentiment models, automation infra, and the experiments that don't make the front page. The receipts are all public.
All repositories ↗I study how people make decisions at the neural level — reward systems, motivation, cognitive bias, impulsivity — then I build systems that account for those realities. My edge isn't that I read Thinking, Fast and Slow; it's that the dopamine circuits and risk-tolerance mechanisms behind trading psychology are literally what I study at the bench.
That's the through-line in everything here: adversarial agents that mirror how good investment committees structure disagreement; content scoring grounded in measured brain response; infrastructure designed to recover without a human in the loop. Syntax is a commodity in 2026 — systems thinking at this level is not.
I'm pursuing an AI product-management track and PMP because the biggest leverage point isn't a better model — it's a better system around people. Long term, I want to build and lead the teams doing that work.
What I shipped, what broke, what I learned — with the diff attached.
Caught a leaked DB credential in a freshly-public repo — rotated first, scrubbed history second, verified green by observation. Secret hygiene is a muscle worth building before you need it.
Pre-registered TRIBE v2's validation protocol before the results — the goalposts live in git history with a timestamp, so future-me can't move them.
Gave my personal memory stack a 3-tier watchdog: detect → auto-heal → only then page me. The page is the last resort, not the first response.
Building decision intelligence, behavioral AI, or agentic tooling? I'd like to compare notes.