Enterprise-grade logistics engine powered by predictive ML and real-time event streaming. Reduce delivery times by 23%.
Live demo runs on free-tier infrastructure
Supabase + Upstash + Railway (fallback-safe, limited workloads)
Full-power stack is local-first
Run locally to unlock Kafka, Flink, Airflow, PostGIS, TimescaleDB, Redis, MongoDB, OSRM.
Built for ops teams: fast reroutes, live visibility, and optimization that feels immediate.
Assign, reroute, and manage exceptions in seconds.
Constraint-based VRP solver with live recalculation.
Forecast load spikes and rebalance capacity early.
Real-time events power the dashboard and alerts.
Built with a modern stack designed for high-throughput data processing and real-time inference.
Prophet + LSTM models triggered via Airflow to predict demand hotspots.
OR-Tools VRP engine re-calculates routes in milliseconds upon new events.
Apache Kafka & Flink handle high-volume telemetry ingestion.
Each capability maps to code in this repo. Local runs show the full stack; the hosted demo stays lightweight.
Integrate with your OMS or TMS using a FastAPI layer that exposes routing, status, and exception handling.
curl -X POST http://localhost:8000/routes/optimize \
-H "Content-Type: application/json" \
-d '{
"depot_id": "hub_12",
"stops": 48,
"constraints": ["capacity", "time_window"]
}'{
"optimized_routes": 12,
"total_miles_saved": 84.5,
"eta_delta_minutes": -18
}