Built by Jimmy Webber
Dynamic Value Chain Engine
Real-time infrastructure intelligence. Cascade simulation. Monte Carlo risk analysis. Distributed AI predicting disruptions across supply chains, construction, and global trade.
Explore the System
Pick a supply chain scenario. Run the cascade. See which nodes fail, how far it propagates, and what to do about it.
Global infrastructure graph with disruption propagation. Real-time risk scores on 117 nodes across ports, energy, data, and trade.
58 sources feeding every hour. Geopolitical events, weather, markets, shipping — watch it update in real time.
Describe what you make or source. The system decomposes it, maps the supply chain, and scores every dependency for risk.
DTC code diagnosis, failure predictions, supplier comparison, and cost estimates for automotive repair shops.
3.7M parameter transformer model. 219 event types across 8 domains. Cross-domain cascade forecasting with swarm enhancement.
Architecture
Prediction Model
6-layer Transformer (PyTorch, MPS). 219 domain-exclusive event types. Trained on 616K real FHWA records + 10K mined live cascades. 51% top-1 accuracy across 278 types.
Simulation Suite
Monte Carlo (6ms), Sensitivity Analysis, Stress Testing, Scenario Trees, Agent-Based competition. 10 programmable variables. Full suite in 18ms.
Data Pipeline
58 sources: NOAA, GDELT, USGS, Finnhub, BLS, RSS. 81K+ events accumulated. FHWA NBI (all 50 states). HPMS pavement. Hourly ingestion.
Infrastructure
4-node swarm (Mac M4 + MSI CUDA + Pi + MacBook). Tailscale mesh. $0/month — zero cloud dependencies. Nightly backups. Dead man switch monitoring.
Python · PyTorch · FastAPI · Next.js · React · Tailwind · Plotly · Tailscale · Cloudflare
DVCE — Built by Jimmy Webber — Solo full-stack AI engineering