Many industries. One connected engine.
Most of these started as separate problems — a wholesale catalog, a trading loop, a Medicaid dataset, a fitness app. Underneath, they run on the same knowledge base and automation stack. Pick an industry to see what I've shipped there, and how it connects to everything else.
12 industries, one stack
Connected by 9 reusable infrastructure threads — auth, multi-tenant data, inventory math, the autoresearch brain, multi-model orchestration. Tap one to dive in.
9 pieces of infrastructure, reused across industries.
The reason one person can ship across this many domains: the hard parts are built once and shared. Each thread below powers every industry listed under it — tap one to jump there.
Multi-model orchestration (Claude / GPT / local)
6×Convex + R2 multi-tenant pattern
5×Autoresearch / shared knowledge engine
5×Clerk auth + tenant isolation
4×Operant-conditioning / reward mechanics
3×Cross-domain synthesis from the knowledge graph
3×Inventory-intelligence engine (WOS, ABC-XYZ, reorder points)
2×Public-data pipelines (EDGAR / CMS / FRED / yfinance)
2×Generative vision (Florence-2 / FLUX / image gen)
2×Counts are generated from the live vault, not hand-typed. “Notes” are Knowledge entries whose subject touches that domain — a single note can touch several industries, so per-industry counts intentionally overlap and don’t sum to the 15,884-note total. “Workflows” are scheduled automations and agents matching each domain.