EcoStream Dispatch — A real-time fleet intelligence platform for climate-aware logistics. It ingests live telemetry, runs sliding-window analytics, detects anomalies (e.g. motor overheat), and surfaces actionable diagnostics.
#FleetManagement#Pathway#StreamingData#DocStore
In EcoStream Dispatch, Pathway RAG powers the diagnostics flow — pulling exact troubleshooting steps from Tata/Mahindra manuals. If RAG is down, it gracefully falls back to DocumentStore.
#RAG#Pathway#FleetTech
What is RAG and how does it work? 🧵
RAG = Retrieval-Augmented Generation.
It works in two steps:
1. Retrieve the most relevant documents (service manuals, PDFs, etc.)
2. Feed those docs + the user query to an LLM
Result: accurate, grounded answers instead of hallucinations.
10/10 What this means for climate impact:
The system is designed to reduce preventable operational waste by detecting high-risk behavior earlier and supporting better dispatch & maintenance timing.
Next step: beta hardening with stronger observability and scale validation.
#RAG
We built EcoStream Dispatch as an alpha real-time fleet intelligence system for climate-aware logistics.
It’s not publicly launched yet, but the architecture is already split like a real product: frontend, API, pipeline worker, docstore, and optional RAG service. #Pathway
9/10 Docker is central to this alpha workflow.
One `docker compose up` runs the full stack reproducibly for demos and team onboarding: frontend, API, pipeline, docstore, and optional RAG.
Realistic end-to-end environment from day one.
2/3 Frontend built with React + Vite + Tailwind.
• Live dashboard with telemetry cards & anomaly alerts
• Fleet table with status & real-time data
• Alert queue with one-click diagnostics
• Clean dark ops-focused UI
Still iterating. Thoughts? #React#Vite#Tailwind